mirror of
https://github.com/openclaw/openclaw.git
synced 2026-06-27 09:52:10 +08:00
Compare commits
1 Commits
feat/comma
...
meow/contr
| Author | SHA1 | Date | |
|---|---|---|---|
|
bf3d17b641 |
@@ -79,7 +79,6 @@ Telegraph style. Root rules only. Read scoped `AGENTS.md` before subtree work.
|
||||
- PR shortlist: `gh pr list ...`; then `gh pr view <n> --json number,title,body,closingIssuesReferences,files,statusCheckRollup,reviewDecision`.
|
||||
- After landing PR: search duplicate open issues/PRs. Before closing: comment why + canonical link.
|
||||
- If an issue/PR is already fixed on current `main` or solved by a new release: comment with proof + canonical commit/PR/release, then close.
|
||||
- `ship` that fixes an issue: after push, comment proof + commit link, then close the issue.
|
||||
- GH comments with markdown backticks, `$`, or shell snippets: avoid inline double-quoted `--body`; use single quotes or `--body-file`.
|
||||
- PR create: description/body always required. Include concise Summary + Verification sections; mention issue/PR refs, behavior changed, and exact local/Testbox/CI proof. Never open an empty-description, empty-body, or placeholder-body PR.
|
||||
- PR execution artifacts/screenshots: attach them to the PR, comment, or an external artifact store. Do not add `.github/pr-assets` or other PR-only assets to the repo.
|
||||
|
||||
18
CHANGELOG.md
18
CHANGELOG.md
@@ -6,35 +6,24 @@ Docs: https://docs.openclaw.ai
|
||||
|
||||
### Changes
|
||||
|
||||
- Auto-reply/queue: prioritize foreground user/manual turns ahead of lower-priority cron, heartbeat, memory, and deferred maintenance work within the same command lane, while preserving FIFO ordering within each priority and promoting old background entries to avoid starvation. Fixes #79589. Thanks @SebTardif.
|
||||
- QA/Mantis: add Telegram live PR evidence automation with Convex-leased credentials, Crabbox transcript capture, motion GIF previews, and inline PR comments.
|
||||
- QA/Mantis: add a Telegram desktop scenario builder that leases Crabbox, installs native Telegram Desktop, configures an OpenClaw Telegram gateway with leased bot credentials, and records VNC screenshot/video artifacts.
|
||||
- Discord/voice: add realtime voice diagnostics for speaker turns, playback resets, barge-in detection, and audio cutoff analysis.
|
||||
- Talk: add `talk.realtime.instructions` so operators can append realtime voice style instructions while preserving OpenClaw's built-in agent-consult guidance. (#79081) Thanks @VACInc.
|
||||
- Discord/voice: default test and source installs to the pure-JS `opusscript` decoder by ignoring optional native `@discordjs/opus` builds, avoiding slow native addon compiles outside dedicated voice-performance lanes.
|
||||
- Discord/voice: add an opt-in native `@discordjs/opus` install script and decoder preference for live voice-performance lanes without charging unrelated Docker/tests for native addon builds.
|
||||
- Gateway/skills: add an opt-in private skill archive upload install path gated by `skills.install.allowUploadedArchives`, so trusted Gateway clients can stage and install zip-backed skills only when operators explicitly enable the code-install surface. (#74430) Thanks @samzong.
|
||||
- Dependencies: refresh workspace pins and patch targets, including ACPX `@agentclientprotocol/claude-agent-acp` `0.33.1`, Codex ACP `0.14.0`, Baileys `7.0.0-rc10`, Google GenAI `2.0.1`, OpenAI `6.37.0`, AWS SDK `3.1045.0`, Kysely `0.29.0`, Tlon skill `0.3.6`, Aimock `1.19.5`, and tsdown `0.22.0`.
|
||||
- Agents/compaction: preserve scoped background exec/process session references across embedded compaction and after-turn runtime contexts without exposing sessions from unrelated scopes. Fixes #79284. (#79307) Thanks @TurboTheTurtle.
|
||||
- Agents/process: tell agents to inspect background sessions with `process log` before sending interactive input and to use `waitingForInput`/`stdinWritable` hints from `log`/`poll`.
|
||||
- CLI/onboarding: improve setup, onboarding, configure, and channel command wayfinding so terminal flows explain the next useful command instead of relying on terse setup labels.
|
||||
- Agents/Codex: remove the configurable Codex dynamic-tools profile so Codex app-server always owns workspace, edit, patch, exec, process, and plan tools while OpenClaw integration tools remain available.
|
||||
|
||||
### Fixes
|
||||
|
||||
- Ollama: stop native `/api/chat` requests from copying catalog `contextWindow` or `maxTokens` into `options.num_ctx` unless `params.num_ctx` is explicitly configured, avoiding pathological prompt-ingestion latency on local large-context models. Fixes #62267. Thanks @BenSHPD.
|
||||
- Ollama: keep the model idle watchdog enabled for `*:cloud` models routed through a local Ollama host, so cloud-backed tool-loop stalls fail over visibly instead of inheriting local-model no-idle behavior. Fixes #79350. Thanks @geek111.
|
||||
- Voice/Ollama: honor routed voice agent `tools.allow` for classic embedded voice responses, including empty allowlists, so no-tool Ollama agents do not receive tool schemas. Fixes #79506. Thanks @donkeykong91.
|
||||
- Gateway: reread config from disk after the first in-process restart loop startup, preventing SIGUSR1 restarts from reusing a stale startup snapshot and dropping config written after boot. Fixes #79947. Thanks @TheLevti.
|
||||
- Codex app-server: deliver native image-generation outputs from Codex `savedPath` events as reply media, so blank-text image generation turns still attach the generated file. Thanks @keshavbotagent.
|
||||
- Network/SSRF: keep pinned automatic DNS lookups on IPv4 when dual-stack hosts also publish AAAA records, and treat `EADDRNOTAVAIL` as a transient gateway network failure instead of a fatal crash. Fixes #80078. Thanks @takamasa-aiso.
|
||||
- Control UI: show compact one-line live/idle/terminal run status badges in the Sessions table and rename the active-minute filter to its updated-within meaning. Fixes #78307. Thanks @BunsDev.
|
||||
- Control UI: scope chat session-list refreshes by agent and skip disk-only agent store discovery for configured-only lists, preventing post-first-message session switching stalls on large Windows stores. Fixes #79675. Thanks @lovelefeng-glitch, @BunsDev.
|
||||
- Control UI: show safe native Codex plan, item, tool, and lifecycle activity in Chat without rendering private reasoning text, so reasoning-enabled Codex runs no longer look stuck behind only the reading indicator. Fixes #80039. Thanks @BunsDev.
|
||||
- Media/host-read: allow buffer-verified gzip, tar, and 7z archives in the shared host-local media validator alongside ZIP and document attachments.
|
||||
- Plugins/doctor: invalidate persisted plugin registry snapshots when plugin diagnostics point at deleted source paths, so `openclaw doctor` stops repeating stale warnings after a local extension is replaced by a managed npm plugin. Fixes #80087. (#80134) Thanks @hclsys.
|
||||
- Cron: let isolated self-cleanup runs inspect their own job run history while keeping other cron jobs and mutation actions blocked. Fixes #80019. Thanks @hclsys.
|
||||
- Cron: report isolated agent-turn setup and pre-model stalls with phase-specific timeout errors instead of waiting for the full job budget when no model call starts. Fixes #74803. Thanks @jeffsteinbok-openclaw and @dgkim311.
|
||||
- CLI/plugins: treat arbitrary unknown subcommands outside plugin CLI metadata as normal unknown commands instead of suggesting `plugins.allow`, while preserving allowlist guidance for real plugin command roots. Fixes #80109. (#80123) Thanks @kagura-agent.
|
||||
- CLI/config: persist explicit `config set` and `config patch` values that equal runtime defaults instead of reporting success while dropping them. Fixes #79856. (#80106) Thanks @abodanty and @hclsys.
|
||||
- OpenAI/realtime voice: accept Codex-compatible legacy audio and transcript event aliases so provider protocol drift does not drop assistant audio or captions.
|
||||
- Discord/voice: keep default agent-proxy realtime sessions from auto-speaking filler before the forced OpenClaw consult answer, finish Discord playback on realtime response completion, and queue later exact-speech answers until playback idles to avoid mid-sentence replacement.
|
||||
@@ -56,7 +45,6 @@ Docs: https://docs.openclaw.ai
|
||||
- Agents: stop blank model-emitted tool calls before dispatch while preserving id-based tool-name recovery, preventing Kimi/NVIDIA blank-name retry loops without creating a callable `_blank` sentinel. Fixes #34129. (#56391) Thanks @smartchainark.
|
||||
- Agents/Telegram: deliver the canonical final assistant answer instead of replaying accumulated pre-tool text blocks, preventing duplicate Telegram replies and raw-looking tool-output fragments from leaking into chat delivery. Fixes #79621 and #79986. Thanks @nonzeroclaw and @dudaefj.
|
||||
- Auto-reply/TUI: keep fallback timeout recovery deliverable after a primary model lifecycle error by emitting fallback progress and deferring terminal TUI errors until recovery has a chance to finish. Fixes #80000. (#80009) Thanks @TurboTheTurtle.
|
||||
- Heartbeat: clear stale auto fallback model overrides when the configured default model changes, so heartbeat runs follow updated `agents.defaults.model.primary` without requiring a manual reset. Fixes #74284. Thanks @brtkwr and @bitloi.
|
||||
- CLI/agent: let `openclaw agent --model` use the backend/admin Gateway scope without cached device-token scopes silently downscoping the request. (#78837) Thanks @VACInc.
|
||||
- CLI/help: keep help and version invocations configless while improving shared port, channel, plugin, task, session, message, pairing, and auth recovery text.
|
||||
- CLI/config: explain strict JSON parse failures with a valid example and the plain-string escape hatch.
|
||||
@@ -77,11 +65,9 @@ Docs: https://docs.openclaw.ai
|
||||
- OpenAI/Codex: point gateway missing-key recovery and wizard docs at the canonical `openai/gpt-5.5` plus Codex OAuth route, and fix trajectory export errors so they suggest the valid `openclaw sessions` command.
|
||||
- Google/Gemini: normalize retired `google/gemini-3-pro-preview` primary, fallback, and model-map refs during config load and unrelated config writes so saved config keeps targeting Gemini 3.1 Pro Preview.
|
||||
- Google/Gemini: normalize retired Gemini 3 Pro Preview ids inside emitted Google provider model config, so regenerated models.json rows test `google/gemini-3.1-pro-preview`.
|
||||
- Google/Gemini: normalize retired Gemini 3 Pro Preview ids preserved from existing merged models.json providers so config emission keeps targeting `google/gemini-3.1-pro-preview`.
|
||||
- GitHub Copilot: mint short-lived Copilot API tokens with the same `vscode-chat` integration identity used by runtime requests, and refresh legacy cached tokens missing that identity so image-capable Copilot models no longer inherit the `copilot-language-server` scope. Fixes #79946, #80074. Thanks @TurboTheTurtle.
|
||||
- Plugins/doctor: drop stale managed npm install records when `openclaw doctor --fix` removes npm packages that shadow bundled plugins, so the rebuilt registry no longer resurrects the removed package metadata.
|
||||
- Discord/voice: reuse or suppress late realtime consult tool calls without stealing newer speaker context or speaking forced fallback answers twice.
|
||||
- Discord/voice: skip likely incomplete realtime forced-consult transcript fragments and non-actionable closings so stale partial speech does not queue delayed answers over the next turn.
|
||||
- Discord/voice: synthesize realtime playback timestamps from emitted Discord PCM so OpenAI realtime barge-in truncation no longer sees `audioEndMs=0` and skips legitimate interruptions.
|
||||
- Plugin SDK: keep activated linked plugin runtime facades loadable when bundled plugin fallback is disabled. Thanks @shakkernerd.
|
||||
- Feishu: auto-thread `message(action="send")` replies inside the topic when the active session is group_topic or group_topic_sender, and propagate `replyInThread` through text, card, and media outbound adapters so topic-scoped sessions no longer post at the group root. Fixes #74903. (#77151) Thanks @ai-hpc.
|
||||
@@ -117,7 +103,6 @@ Docs: https://docs.openclaw.ai
|
||||
### Changes
|
||||
|
||||
- Skills: add `skills.load.allowSymlinkTargets` so intentional symlinked skill folders can resolve into trusted sibling repos without disabling root containment.
|
||||
- Agents/tools: add core Tool Search so agents can search and call large OpenClaw, MCP, and client tool catalogs through one compact PI bridge.
|
||||
- Chat commands: add `/think default` and `/fast default` to clear session overrides and inherit configured/provider defaults. (#79385) Thanks @VACInc.
|
||||
- Dependencies: refresh workspace dependency pins and lockfile, including `@openai/codex` `0.130.0`, `acpx` `0.7.0`, AWS SDK `3.1044.0`, OpenTelemetry `0.217.0`, `typebox` `1.1.38`, `vite` `8.0.11`, `oxfmt` `0.48.0`, and `oxlint` `1.63.0`, and update the Codex harness model snapshot for the new bundled app-server catalog.
|
||||
- Plugins/install: add guarded plugin install overrides so onboarding and repair tests can route specific plugins to registry specs or local `npm pack` artifacts via environment variables.
|
||||
@@ -193,7 +178,6 @@ Docs: https://docs.openclaw.ai
|
||||
- Matrix: attach `com.openclaw.presentation` metadata to semantic presentation replies so OpenClaw-aware Matrix clients can render rich buttons, selects, context rows, and dividers while stock clients keep the plain text fallback. (#73312) Thanks @kakahu2015.
|
||||
- Codex app-server: disarm the short post-tool completion watchdog after current-turn activity, expose `appServer.turnCompletionIdleTimeoutMs`, and include raw assistant item context in idle-timeout diagnostics so status-only post-tool stalls stop failing as idle. Fixes #77984. Thanks @roseware-dev and @rubencu.
|
||||
- Plugin skills/Windows: publish plugin-provided skill directories as junctions on Windows so standard users without Developer Mode can register plugin skills without symlink EPERM failures. Fixes #77958. (#77971) Thanks @hclsys and @jarro.
|
||||
- Process tool: show input-wait hints from `log` and `poll` for idle interactive background sessions so operators can inspect stuck CLIs and resume them with existing input actions. Fixes #33957. Thanks @bitloi and @vincentkoc.
|
||||
- Shell env/Windows: hide the login-shell environment probe child window so gateway startup and shell-env refreshes do not flash a console on Windows. Fixes #78159. (#78266) Thanks @BradGroux.
|
||||
- MS Teams: surface blocked Bot Framework egress by logging JWKS fetch network failures and adding a Bot Connector send hint for transport-level reply failures. Fixes #77674. (#78081) Thanks @Beandon13.
|
||||
- Media/host-read: allow buffer-verified ZIP archives in the host-local media validator so agents can send ZIP attachments via the message tool. Fixes #78057. (#78292) Thanks @Linux2010.
|
||||
|
||||
@@ -716,7 +716,6 @@ public struct AgentParams: Codable, Sendable {
|
||||
public let bootstrapcontextmode: AnyCodable?
|
||||
public let bootstrapcontextrunkind: AnyCodable?
|
||||
public let acpturnsource: String?
|
||||
public let internalruntimehandoffid: String?
|
||||
public let internalevents: [[String: AnyCodable]]?
|
||||
public let inputprovenance: [String: AnyCodable]?
|
||||
public let voicewaketrigger: String?
|
||||
@@ -753,7 +752,6 @@ public struct AgentParams: Codable, Sendable {
|
||||
bootstrapcontextmode: AnyCodable?,
|
||||
bootstrapcontextrunkind: AnyCodable?,
|
||||
acpturnsource: String?,
|
||||
internalruntimehandoffid: String?,
|
||||
internalevents: [[String: AnyCodable]]?,
|
||||
inputprovenance: [String: AnyCodable]?,
|
||||
voicewaketrigger: String?,
|
||||
@@ -789,7 +787,6 @@ public struct AgentParams: Codable, Sendable {
|
||||
self.bootstrapcontextmode = bootstrapcontextmode
|
||||
self.bootstrapcontextrunkind = bootstrapcontextrunkind
|
||||
self.acpturnsource = acpturnsource
|
||||
self.internalruntimehandoffid = internalruntimehandoffid
|
||||
self.internalevents = internalevents
|
||||
self.inputprovenance = inputprovenance
|
||||
self.voicewaketrigger = voicewaketrigger
|
||||
@@ -827,7 +824,6 @@ public struct AgentParams: Codable, Sendable {
|
||||
case bootstrapcontextmode = "bootstrapContextMode"
|
||||
case bootstrapcontextrunkind = "bootstrapContextRunKind"
|
||||
case acpturnsource = "acpTurnSource"
|
||||
case internalruntimehandoffid = "internalRuntimeHandoffId"
|
||||
case internalevents = "internalEvents"
|
||||
case inputprovenance = "inputProvenance"
|
||||
case voicewaketrigger = "voiceWakeTrigger"
|
||||
|
||||
@@ -862,25 +862,5 @@
|
||||
{
|
||||
"source": "fs-safe Cleanup Plan",
|
||||
"target": "fs-safe Cleanup Plan"
|
||||
},
|
||||
{
|
||||
"source": "Tool Search",
|
||||
"target": "工具搜索"
|
||||
},
|
||||
{
|
||||
"source": "Tools and plugins",
|
||||
"target": "工具和插件"
|
||||
},
|
||||
{
|
||||
"source": "Multi-agent sandbox and tools",
|
||||
"target": "多 Agent 沙盒和工具"
|
||||
},
|
||||
{
|
||||
"source": "Exec tool",
|
||||
"target": "Exec 工具"
|
||||
},
|
||||
{
|
||||
"source": "ACP agents setup",
|
||||
"target": "ACP Agents 设置"
|
||||
}
|
||||
]
|
||||
|
||||
@@ -1214,7 +1214,7 @@ Notes:
|
||||
- If `voice.autoJoin` has multiple entries for the same guild, OpenClaw joins the last configured channel for that guild.
|
||||
- `voice.daveEncryption` and `voice.decryptionFailureTolerance` pass through to `@discordjs/voice` join options.
|
||||
- `@discordjs/voice` defaults are `daveEncryption=true` and `decryptionFailureTolerance=24` if unset.
|
||||
- OpenClaw defaults to the pure-JS `opusscript` decoder for Discord voice receive. The optional native `@discordjs/opus` package is ignored by the repo pnpm install policy so normal installs, Docker lanes, and unrelated tests do not compile a native addon. Dedicated voice-performance hosts can opt in with `OPENCLAW_DISCORD_OPUS_DECODER=native` after installing the native addon.
|
||||
- OpenClaw defaults to the pure-JS `opusscript` decoder for Discord voice receive. The optional native `@discordjs/opus` package is ignored by the repo pnpm install policy so normal installs and tests do not compile a native addon; only opt into a native opus build in a dedicated voice-performance or live-lane environment.
|
||||
- `voice.connectTimeoutMs` controls the initial `@discordjs/voice` Ready wait for `/vc join` and auto-join attempts. Default: `30000`.
|
||||
- `voice.reconnectGraceMs` controls how long OpenClaw waits for a disconnected voice session to begin reconnecting before destroying it. Default: `15000`.
|
||||
- In `stt-tts` mode, voice playback does not stop just because another user starts speaking. To avoid feedback loops, OpenClaw ignores new voice capture while TTS is playing; speak after playback finishes for the next turn. Realtime modes forward speaker starts as barge-in signals to the realtime provider.
|
||||
@@ -1225,24 +1225,6 @@ Notes:
|
||||
- If receive logs repeatedly show `DecryptionFailed(UnencryptedWhenPassthroughDisabled)` after updating, collect a dependency report and logs. The bundled `@discordjs/voice` line includes the upstream padding fix from discord.js PR #11449, which closed discord.js issue #11419.
|
||||
- `The operation was aborted` receive events are expected when OpenClaw finalizes a captured speaker segment; they are verbose diagnostics, not warnings.
|
||||
- Verbose Discord voice logs include a bounded one-line STT transcript preview for each accepted speaker segment, so debugging shows both the user side and the agent reply side without dumping unbounded transcript text.
|
||||
- In `agent-proxy` mode, forced consult fallback skips likely incomplete transcript fragments such as text ending in `...` or a trailing connector like `and`, plus obvious non-actionable closings like “be right back” or “bye”. Logs show `forced agent consult skipped reason=...` when this prevents a stale queued answer.
|
||||
|
||||
Native opus setup for source checkouts:
|
||||
|
||||
```bash
|
||||
pnpm install
|
||||
mise exec node@22 -- pnpm discord:opus:install
|
||||
```
|
||||
|
||||
Use Node 22 for the gateway when you want the upstream macOS arm64 prebuilt native addon. If you use another Node runtime, the opt-in installer may need a local `node-gyp` source-build toolchain.
|
||||
|
||||
After installing the native addon, start the Gateway with:
|
||||
|
||||
```bash
|
||||
OPENCLAW_DISCORD_OPUS_DECODER=native pnpm gateway:watch
|
||||
```
|
||||
|
||||
Verbose voice logs should show `discord voice: opus decoder: @discordjs/opus`. Without the env opt-in, or if the native addon is missing or cannot load on the host, OpenClaw logs `discord voice: opus decoder: opusscript` and keeps receiving voice through the pure-JS fallback.
|
||||
|
||||
STT plus TTS pipeline:
|
||||
|
||||
@@ -1389,7 +1371,6 @@ Expected voice logs:
|
||||
- On join: `discord voice: joining ... voiceSession=... supervisorSession=... agentSessionMode=... voiceModel=... realtimeModel=...`
|
||||
- On realtime start: `discord voice: realtime bridge starting ... autoRespond=false interruptResponse=false bargeIn=false minBargeInAudioEndMs=...`
|
||||
- On speaker audio: `discord voice: realtime speaker turn opened ...`, `discord voice: realtime input audio started ... outputAudioMs=... outputActive=...`, and `discord voice: realtime speaker turn closed ... chunks=... discordBytes=... realtimeBytes=... interruptedPlayback=...`
|
||||
- On skipped stale speech: `discord voice: realtime forced agent consult skipped reason=incomplete-transcript ...` or `reason=non-actionable-closing ...`
|
||||
- On realtime response completion: `discord voice: realtime audio playback finishing reason=response.done ... audioMs=... chunks=...`
|
||||
- On playback stop/reset: `discord voice: realtime audio playback stopped reason=... audioMs=... elapsedMs=... chunks=...`
|
||||
- On realtime consult: `discord voice: realtime consult requested ... voiceSession=... supervisorSession=... question=...`
|
||||
|
||||
@@ -60,7 +60,7 @@ OpenClaw separates the selected provider/model from why it was selected. That so
|
||||
|
||||
- **Configured default**: `agents.defaults.model.primary` uses `agents.defaults.model.fallbacks`.
|
||||
- **Agent primary**: `agents.list[].model` is strict unless that agent model object includes its own `fallbacks`. Use `fallbacks: []` to make the strict behavior explicit, or provide a non-empty list to opt that agent into model fallback.
|
||||
- **Auto fallback override**: a runtime fallback writes `providerOverride`, `modelOverride`, `modelOverrideSource: "auto"`, and the selected origin model before retrying. That auto override can keep walking the configured fallback chain and is cleared by `/new`, `/reset`, and `sessions.reset`. Heartbeat runs without an explicit `heartbeat.model` also clear a direct auto override when its origin no longer matches the current configured default.
|
||||
- **Auto fallback override**: a runtime fallback writes `providerOverride`, `modelOverride`, and `modelOverrideSource: "auto"` before retrying. That auto override can keep walking the configured fallback chain and is cleared by `/new`, `/reset`, and `sessions.reset`.
|
||||
- **User session override**: `/model`, the model picker, `session_status(model=...)`, and `sessions.patch` write `modelOverrideSource: "user"`. That is an exact session selection. If the selected provider/model fails before producing a reply, OpenClaw reports the failure instead of answering from an unrelated configured fallback.
|
||||
- **Legacy session override**: older session entries may have `modelOverride` without `modelOverrideSource`. OpenClaw treats those as user overrides so an explicit old selection is not silently converted into fallback behavior.
|
||||
- **Cron payload model**: a cron job `payload.model` / `--model` is a job primary, not a user session override. It uses configured fallbacks unless the job provides `payload.fallbacks`; `payload.fallbacks: []` makes the cron run strict.
|
||||
|
||||
@@ -15,10 +15,9 @@ We serialize inbound auto-reply runs (all channels) through a tiny in-process qu
|
||||
|
||||
## How it works
|
||||
|
||||
- A lane-aware queue drains each lane with a configurable concurrency cap (default 1 for unconfigured lanes; main defaults to 4, subagent to 8). Entries with the same priority remain FIFO; user/manual turns can jump ahead of lower-priority background work in the same lane.
|
||||
- A lane-aware FIFO queue drains each lane with a configurable concurrency cap (default 1 for unconfigured lanes; main defaults to 4, subagent to 8).
|
||||
- `runEmbeddedPiAgent` enqueues by **session key** (lane `session:<key>`) to guarantee only one active run per session.
|
||||
- Each session run is then queued into a **global lane** (`main` by default) so overall parallelism is capped by `agents.defaults.maxConcurrent`.
|
||||
- Priority is local to a lane. It does not interrupt an active run; it only chooses the next queued entry when a lane has capacity. A starvation guard promotes old low/normal-priority entries after a wait threshold.
|
||||
- When verbose logging is enabled, queued runs emit a short notice if they waited more than ~2s before starting.
|
||||
- Typing indicators still fire immediately on enqueue (when supported by the channel) so user experience is unchanged while we wait our turn.
|
||||
|
||||
|
||||
@@ -1280,7 +1280,6 @@
|
||||
"tools/reactions",
|
||||
"tools/thinking",
|
||||
"tools/tokenjuice",
|
||||
"tools/tool-search",
|
||||
"tools/loop-detection",
|
||||
"tools/trajectory",
|
||||
"tools/tts",
|
||||
|
||||
@@ -46,7 +46,6 @@ Environment overrides:
|
||||
- `PI_BASH_MAX_OUTPUT_CHARS`: in-memory output cap (chars)
|
||||
- `OPENCLAW_BASH_PENDING_MAX_OUTPUT_CHARS`: pending stdout/stderr cap per stream (chars)
|
||||
- `PI_BASH_JOB_TTL_MS`: TTL for finished sessions (ms, bounded to 1m–3h)
|
||||
- `OPENCLAW_PROCESS_INPUT_WAIT_IDLE_MS`: idle-output threshold before writable background sessions are marked as likely waiting for input (default 15000 ms)
|
||||
|
||||
Config (preferred):
|
||||
|
||||
@@ -62,7 +61,7 @@ Actions:
|
||||
|
||||
- `list`: running + finished sessions
|
||||
- `poll`: drain new output for a session (also reports exit status)
|
||||
- `log`: read the aggregated output and show input recovery hints (supports `offset` + `limit`)
|
||||
- `log`: read the aggregated output (supports `offset` + `limit`)
|
||||
- `write`: send stdin (`data`, optional `eof`)
|
||||
- `send-keys`: send explicit key tokens or bytes to a PTY-backed session
|
||||
- `submit`: send Enter / carriage return to a PTY-backed session
|
||||
@@ -79,14 +78,9 @@ Notes:
|
||||
- `process` is scoped per agent; it only sees sessions started by that agent.
|
||||
- Use `poll` / `log` for status, logs, quiet-success confirmation, or
|
||||
completion confirmation when automatic completion wake is unavailable.
|
||||
- Use `log` before recovering an interactive CLI so the current transcript,
|
||||
stdin state, and input-wait hint are visible together.
|
||||
- Use `write` / `send-keys` / `submit` / `paste` / `kill` when you need input
|
||||
or intervention.
|
||||
- `process list` includes a derived `name` (command verb + target) for quick scans.
|
||||
- `process list`, `poll`, and `log` report `waitingForInput` only
|
||||
when the session still has writable stdin and has been idle longer than the
|
||||
input-wait threshold.
|
||||
- `process log` uses line-based `offset`/`limit`.
|
||||
- When both `offset` and `limit` are omitted, it returns the last 200 lines and includes a paging hint.
|
||||
- When `offset` is provided and `limit` is omitted, it returns from `offset` to the end (not capped to 200).
|
||||
@@ -105,12 +99,6 @@ Run a long task and poll later:
|
||||
{ "tool": "process", "action": "poll", "sessionId": "<id>" }
|
||||
```
|
||||
|
||||
Inspect an interactive session before sending input:
|
||||
|
||||
```json
|
||||
{ "tool": "process", "action": "log", "sessionId": "<id>" }
|
||||
```
|
||||
|
||||
Start immediately in background:
|
||||
|
||||
```json
|
||||
|
||||
@@ -200,9 +200,6 @@ settings such as `agents.defaults.imageGenerationModel`, `videoGenerationModel`,
|
||||
|
||||
Text, images, video, music, TTS, approvals, and messaging-tool output continue
|
||||
through the normal OpenClaw delivery path. Media generation does not require PI.
|
||||
When Codex emits a native image-generation item with a `savedPath`, OpenClaw
|
||||
forwards that exact file through the normal reply-media path even if the Codex
|
||||
turn has no assistant text.
|
||||
|
||||
## Related
|
||||
|
||||
|
||||
@@ -117,13 +117,6 @@ tool descriptor during discovery and caches it by plugin source and contract, so
|
||||
later tool planning can skip plugin runtime loading. Tool execution still loads
|
||||
the owning plugin and calls the live registered implementation.
|
||||
|
||||
[Tool Search](/tools/tool-search) is the compact surface
|
||||
for large catalogs. Instead of putting every OpenClaw, MCP, or client tool
|
||||
schema into the prompt, OpenClaw can give the model an isolated Node runtime
|
||||
with `openclaw.tools.search`, `openclaw.tools.describe`, and
|
||||
`openclaw.tools.call`. Calls still flow back through the Gateway, so tool
|
||||
policy, approvals, hooks, and session logs remain authoritative.
|
||||
|
||||
## Tool configuration
|
||||
|
||||
### Allow and deny lists
|
||||
|
||||
@@ -1,260 +0,0 @@
|
||||
---
|
||||
summary: "Tool Search: compact large PI tool catalogs behind search, describe, and call"
|
||||
title: "Tool Search"
|
||||
read_when:
|
||||
- You want PI agents to use a large tool catalog without adding every tool schema to the prompt
|
||||
- You want OpenClaw tools, MCP tools, and client tools exposed through one compact PI surface
|
||||
- You are implementing or debugging tool discovery for PI runs
|
||||
---
|
||||
|
||||
Tool Search gives PI agents one compact way to discover and call large tool
|
||||
catalogs. It is useful when the run has many available tools but the model is
|
||||
likely to need only a few of them.
|
||||
|
||||
When enabled for PI, the model receives one `tool_search_code` tool by default.
|
||||
That tool runs a short JavaScript body in an isolated Node subprocess with an
|
||||
`openclaw.tools` bridge:
|
||||
|
||||
```js
|
||||
const hits = await openclaw.tools.search("create a GitHub issue");
|
||||
const tool = await openclaw.tools.describe(hits[0].id);
|
||||
return await openclaw.tools.call(tool.id, {
|
||||
title: "Crash on startup",
|
||||
body: "Steps to reproduce...",
|
||||
});
|
||||
```
|
||||
|
||||
The catalog can include OpenClaw tools, plugin tools, MCP tools, and
|
||||
client-provided tools. The model does not see every full schema up front.
|
||||
Instead, it searches compact descriptors, describes one selected tool when it
|
||||
needs the exact schema, and calls that tool through OpenClaw.
|
||||
|
||||
Codex harness runs do not receive these OpenClaw Tool Search controls. OpenClaw
|
||||
passes product capabilities to Codex as dynamic tools, and Codex owns native
|
||||
code mode, native tool search, deferred dynamic tools, and nested tool calls.
|
||||
|
||||
## How a turn runs
|
||||
|
||||
At planning time the PI embedded runner builds the effective catalog for the
|
||||
run:
|
||||
|
||||
1. Resolve the active tool policy for the agent, profile, sandbox, and session.
|
||||
2. List eligible OpenClaw and plugin tools.
|
||||
3. List eligible MCP tools through the session MCP runtime.
|
||||
4. Add eligible client tools supplied for the current run.
|
||||
5. Index compact descriptors for search.
|
||||
6. Expose either the PI code bridge or the structured fallback tools to the
|
||||
model.
|
||||
|
||||
At execution time every real tool call returns to OpenClaw. The isolated Node
|
||||
runtime does not hold plugin implementations, MCP client objects, or secrets.
|
||||
`openclaw.tools.call(...)` crosses the bridge back into the Gateway, where the
|
||||
normal policy, approval, hook, logging, and result handling still apply.
|
||||
|
||||
## Modes
|
||||
|
||||
`tools.toolSearch` has two model-facing modes:
|
||||
|
||||
- `code`: exposes `tool_search_code`, the default compact JavaScript bridge.
|
||||
- `tools`: exposes `tool_search`, `tool_describe`, and `tool_call` as plain
|
||||
structured tools for providers that should not receive code.
|
||||
|
||||
Both modes use the same catalog and execution path. The only difference is the
|
||||
shape the model sees. If the current runtime cannot launch the isolated Node
|
||||
code-mode child process, the default `code` mode falls back to `tools` before
|
||||
catalog compaction.
|
||||
|
||||
There is no separate source-selection config. When Tool Search is enabled, the
|
||||
catalog includes eligible OpenClaw, MCP, and client tools after normal policy
|
||||
filtering.
|
||||
|
||||
## Why this exists
|
||||
|
||||
Large catalogs are useful but expensive. Sending every tool schema to the model
|
||||
makes the request larger, slows planning, and increases accidental tool
|
||||
selection.
|
||||
|
||||
Tool Search changes the shape:
|
||||
|
||||
- direct tools: the model sees every selected schema before the first token
|
||||
- Tool Search code mode: the model sees one compact code tool and a short API
|
||||
contract
|
||||
- Tool Search tools mode: the model sees three compact structured fallback
|
||||
tools
|
||||
- during the turn: the model loads only the tool schemas it actually needs
|
||||
|
||||
Direct tool exposure is still the right default for small catalogs. Tool Search
|
||||
is best when one run can see many tools, especially from MCP servers or
|
||||
client-provided app tools.
|
||||
|
||||
## API
|
||||
|
||||
`openclaw.tools.search(query, options?)`
|
||||
|
||||
Searches the effective catalog for the current run. Results are compact and safe
|
||||
to put back into prompt context.
|
||||
|
||||
```js
|
||||
const hits = await openclaw.tools.search("calendar event", { limit: 5 });
|
||||
```
|
||||
|
||||
`openclaw.tools.describe(id)`
|
||||
|
||||
Loads full metadata for one search result, including the exact input schema.
|
||||
|
||||
```js
|
||||
const calendarCreate = await openclaw.tools.describe("mcp:calendar:create_event");
|
||||
```
|
||||
|
||||
`openclaw.tools.call(id, args)`
|
||||
|
||||
Calls a selected tool through OpenClaw.
|
||||
|
||||
```js
|
||||
await openclaw.tools.call(calendarCreate.id, {
|
||||
summary: "Planning",
|
||||
start: "2026-05-09T14:00:00Z",
|
||||
});
|
||||
```
|
||||
|
||||
The structured fallback mode exposes the same operations as tools:
|
||||
|
||||
- `tool_search`
|
||||
- `tool_describe`
|
||||
- `tool_call`
|
||||
|
||||
## Runtime boundary
|
||||
|
||||
The code bridge runs in a short-lived Node subprocess. The subprocess starts
|
||||
with Node permission mode enabled, an empty environment, no filesystem or
|
||||
network grants, and no child-process or worker grants. OpenClaw enforces a
|
||||
parent-process wall-clock timeout and kills the subprocess on timeout, including
|
||||
after async continuations.
|
||||
|
||||
The runtime exposes only:
|
||||
|
||||
- `console.log`, `console.warn`, and `console.error`
|
||||
- `openclaw.tools.search`
|
||||
- `openclaw.tools.describe`
|
||||
- `openclaw.tools.call`
|
||||
|
||||
Normal OpenClaw behavior still applies to final calls:
|
||||
|
||||
- tool allow and deny policies
|
||||
- per-agent and per-sandbox tool restrictions
|
||||
- owner-only gating
|
||||
- approval hooks
|
||||
- plugin `before_tool_call` hooks
|
||||
- session identity, logs, and telemetry
|
||||
|
||||
## Config
|
||||
|
||||
Enable Tool Search for PI runs with the default code bridge:
|
||||
|
||||
```bash
|
||||
openclaw config set tools.toolSearch true
|
||||
```
|
||||
|
||||
Equivalent JSON:
|
||||
|
||||
```json5
|
||||
{
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
}
|
||||
```
|
||||
|
||||
Use the structured fallback tools instead for PI runs:
|
||||
|
||||
```json5
|
||||
{
|
||||
tools: {
|
||||
toolSearch: {
|
||||
mode: "tools",
|
||||
},
|
||||
},
|
||||
}
|
||||
```
|
||||
|
||||
Tune code-mode timeout and search result limits:
|
||||
|
||||
```json5
|
||||
{
|
||||
tools: {
|
||||
toolSearch: {
|
||||
mode: "code",
|
||||
codeTimeoutMs: 10000,
|
||||
searchDefaultLimit: 8,
|
||||
maxSearchLimit: 20,
|
||||
},
|
||||
},
|
||||
}
|
||||
```
|
||||
|
||||
Disable it:
|
||||
|
||||
```json5
|
||||
{
|
||||
tools: {
|
||||
toolSearch: false,
|
||||
},
|
||||
}
|
||||
```
|
||||
|
||||
## Prompt and telemetry
|
||||
|
||||
Tool Search records enough telemetry to compare it with direct tool exposure:
|
||||
|
||||
- total serialized tool and prompt bytes sent to the harness
|
||||
- catalog size and source breakdown
|
||||
- search, describe, and call counts
|
||||
- final tool calls executed through OpenClaw
|
||||
- selected tool ids and sources
|
||||
|
||||
Session logs should make it possible to answer:
|
||||
|
||||
- how many tool schemas the model saw up front
|
||||
- how many search and describe operations it performed
|
||||
- which final tool was called
|
||||
- whether the result came from OpenClaw, MCP, or a client tool
|
||||
|
||||
## E2E validation
|
||||
|
||||
The gateway E2E runner proves both paths with the PI harness:
|
||||
|
||||
```bash
|
||||
node --import tsx scripts/tool-search-gateway-e2e.ts
|
||||
```
|
||||
|
||||
It creates a temporary fake plugin with a large tool catalog, starts the mock
|
||||
OpenAI provider, starts a Gateway once in direct mode and once with Tool Search
|
||||
enabled, then compares provider request payloads and session logs.
|
||||
|
||||
The regression proves:
|
||||
|
||||
1. Direct mode can call the fake plugin tool.
|
||||
2. Tool Search can call the same fake plugin tool.
|
||||
3. Direct mode exposes the fake plugin tool schemas directly to the provider.
|
||||
4. Tool Search exposes only the compact bridge.
|
||||
5. The Tool Search request payload is smaller for the large fake catalog.
|
||||
6. Session logs show the expected tool-call counts and bridged call telemetry.
|
||||
|
||||
## Failure behavior
|
||||
|
||||
Tool Search should fail closed:
|
||||
|
||||
- if a tool is not in the effective policy, search should not return it
|
||||
- if a selected tool becomes unavailable, `tool_call` should fail
|
||||
- if policy or approval blocks execution, the call result should report that
|
||||
block instead of bypassing it
|
||||
- if the code bridge cannot create an isolated runtime, use `mode: "tools"` or
|
||||
disable Tool Search for that deployment
|
||||
|
||||
## Related
|
||||
|
||||
- [Tools and plugins](/tools)
|
||||
- [Multi-agent sandbox and tools](/tools/multi-agent-sandbox-tools)
|
||||
- [Exec tool](/tools/exec)
|
||||
- [ACP agents setup](/tools/acp-agents-setup)
|
||||
- [Building plugins](/plugins/building-plugins)
|
||||
@@ -99,6 +99,7 @@ Imported themes are stored only in the current browser profile. They are not wri
|
||||
- Chat history refreshes request a bounded recent window with per-message text caps so large sessions do not force the browser to render a full transcript payload before the chat becomes usable.
|
||||
- Talk through browser realtime sessions. OpenAI uses direct WebRTC, Google Live uses a constrained one-use browser token over WebSocket, and backend-only realtime voice plugins use the Gateway relay transport. Client-owned provider sessions start with `talk.client.create`; Gateway relay sessions start with `talk.session.create`. The relay keeps provider credentials on the Gateway while the browser streams microphone PCM through `talk.session.appendAudio` and forwards `openclaw_agent_consult` provider tool calls through `talk.client.toolCall` for Gateway policy and the larger configured OpenClaw model.
|
||||
- Stream tool calls + live tool output cards in Chat (agent events).
|
||||
- Native Codex runs show a compact safe activity indicator for plan, item, tool, and lifecycle progress. The Control UI does not render raw private reasoning text from Codex app-server events; when reasoning visibility is enabled, the indicator only shows approved progress summaries.
|
||||
|
||||
</Accordion>
|
||||
<Accordion title="Channels, instances, sessions, dreams">
|
||||
|
||||
@@ -12,51 +12,6 @@ vi.mock("openclaw/plugin-sdk/agent-harness-runtime", async (importOriginal) => (
|
||||
|
||||
const mockCallGatewayTool = vi.mocked(callGatewayTool);
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function gatewayRequestPayload(callIndex = 0) {
|
||||
return requireRecord(
|
||||
mockCallGatewayTool.mock.calls[callIndex]?.[2],
|
||||
`gateway request payload ${callIndex + 1}`,
|
||||
);
|
||||
}
|
||||
|
||||
function gatewayCallOptions(callIndex = 0) {
|
||||
return mockCallGatewayTool.mock.calls[callIndex]?.[3];
|
||||
}
|
||||
|
||||
function findApprovalEvent(
|
||||
params: EmbeddedRunAttemptParams,
|
||||
fields: { status?: string; approvalId?: string; command?: string; reason?: string },
|
||||
) {
|
||||
const onAgentEvent = params.onAgentEvent as unknown as { mock?: { calls?: unknown[][] } };
|
||||
const calls = onAgentEvent.mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error("Expected onAgentEvent mock calls");
|
||||
}
|
||||
for (const call of calls) {
|
||||
const event = requireRecord(call[0], "agent event");
|
||||
if (event.stream !== "approval") {
|
||||
continue;
|
||||
}
|
||||
const data = requireRecord(event.data, "approval event data");
|
||||
if (
|
||||
(!fields.status || data.status === fields.status) &&
|
||||
(!fields.approvalId || data.approvalId === fields.approvalId) &&
|
||||
(!fields.command || data.command === fields.command) &&
|
||||
(!fields.reason || data.reason === fields.reason)
|
||||
) {
|
||||
return data;
|
||||
}
|
||||
}
|
||||
throw new Error(`Expected approval event ${JSON.stringify(fields)}`);
|
||||
}
|
||||
|
||||
function createParams(): EmbeddedRunAttemptParams {
|
||||
return {
|
||||
sessionKey: "agent:main:session-1",
|
||||
@@ -98,21 +53,30 @@ describe("Codex app-server approval bridge", () => {
|
||||
"plugin.approval.request",
|
||||
"plugin.approval.waitDecision",
|
||||
]);
|
||||
expect(mockCallGatewayTool.mock.calls[0]?.[0]).toBe("plugin.approval.request");
|
||||
expect(typeof mockCallGatewayTool.mock.calls[0]?.[1]).toBe("object");
|
||||
const requestPayload = gatewayRequestPayload();
|
||||
expect(requestPayload.pluginId).toBe("openclaw-codex-app-server");
|
||||
expect(requestPayload.title).toBe("Codex app-server command approval");
|
||||
expect(requestPayload.twoPhase).toBe(true);
|
||||
expect(requestPayload.turnSourceChannel).toBe("telegram");
|
||||
expect(requestPayload.turnSourceTo).toBe("chat-1");
|
||||
expect(gatewayCallOptions()).toEqual({ expectFinal: false });
|
||||
expect(
|
||||
findApprovalEvent(params, { status: "pending", approvalId: "plugin:approval-1" }),
|
||||
).toBeDefined();
|
||||
expect(
|
||||
findApprovalEvent(params, { status: "approved", approvalId: "plugin:approval-1" }),
|
||||
).toBeDefined();
|
||||
expect(mockCallGatewayTool).toHaveBeenCalledWith(
|
||||
"plugin.approval.request",
|
||||
expect.any(Object),
|
||||
expect.objectContaining({
|
||||
pluginId: "openclaw-codex-app-server",
|
||||
title: "Codex app-server command approval",
|
||||
twoPhase: true,
|
||||
turnSourceChannel: "telegram",
|
||||
turnSourceTo: "chat-1",
|
||||
}),
|
||||
{ expectFinal: false },
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({ status: "pending", approvalId: "plugin:approval-1" }),
|
||||
}),
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({ status: "approved", approvalId: "plugin:approval-1" }),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("describes command approvals from parsed command actions when available", async () => {
|
||||
@@ -135,10 +99,23 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
const requestPayload = gatewayRequestPayload();
|
||||
expect(String(requestPayload.description)).toContain("Command: pnpm test extensions/codex");
|
||||
expect(String(requestPayload.description)).not.toContain("bash -lc");
|
||||
expect(findApprovalEvent(params, { command: "pnpm test extensions/codex" })).toBeDefined();
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: expect.stringContaining("Command: pnpm test extensions/codex"),
|
||||
}),
|
||||
);
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: expect.not.stringContaining("bash -lc"),
|
||||
}),
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({ command: "pnpm test extensions/codex" }),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("describes command approval permission and policy amendments", async () => {
|
||||
@@ -237,15 +214,22 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
expect(gatewayRequestPayload().description).toBe(
|
||||
"Command: pnpm test --watch extensions/codex/src/app-server\nSession: agent:main:session-1",
|
||||
);
|
||||
expect(
|
||||
findApprovalEvent(params, {
|
||||
status: "pending",
|
||||
command: "pnpm test --watch extensions/codex/src/app-server",
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description:
|
||||
"Command: pnpm test --watch extensions/codex/src/app-server\nSession: agent:main:session-1",
|
||||
}),
|
||||
).toBeDefined();
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({
|
||||
status: "pending",
|
||||
command: "pnpm test --watch extensions/codex/src/app-server",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("escapes command approval previews before forwarding approval text and events", async () => {
|
||||
@@ -275,12 +259,15 @@ describe("Codex app-server approval bridge", () => {
|
||||
expect(description).not.toContain("<@U123>");
|
||||
expect(description).not.toContain("[trusted](https://evil)");
|
||||
expect(description).not.toContain("@here");
|
||||
expect(
|
||||
findApprovalEvent(params, {
|
||||
command:
|
||||
"printf '<\uff20U123> \uff3btrusted\uff3d\uff08https://evil\uff09 \uff20here'",
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({
|
||||
command:
|
||||
"printf '<\uff20U123> \uff3btrusted\uff3d\uff08https://evil\uff09 \uff20here'",
|
||||
}),
|
||||
}),
|
||||
).toBeDefined();
|
||||
);
|
||||
});
|
||||
|
||||
it("preserves visible OSC-8 link labels in command previews", async () => {
|
||||
@@ -303,10 +290,18 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
expect(gatewayRequestPayload().description).toBe(
|
||||
"Command: prefix VISIBLE suffix\nSession: agent:main:session-1",
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: "Command: prefix VISIBLE suffix\nSession: agent:main:session-1",
|
||||
}),
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({ command: "prefix VISIBLE suffix" }),
|
||||
}),
|
||||
);
|
||||
expect(findApprovalEvent(params, { command: "prefix VISIBLE suffix" })).toBeDefined();
|
||||
});
|
||||
|
||||
it("strips bidi and invisible formatting controls from command previews", async () => {
|
||||
@@ -328,10 +323,18 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
expect(gatewayRequestPayload().description).toBe(
|
||||
"Command: echo safe cod.exe hidden done\nSession: agent:main:session-1",
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: "Command: echo safe cod.exe hidden done\nSession: agent:main:session-1",
|
||||
}),
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({ command: "echo safe cod.exe hidden done" }),
|
||||
}),
|
||||
);
|
||||
expect(findApprovalEvent(params, { command: "echo safe cod.exe hidden done" })).toBeDefined();
|
||||
});
|
||||
|
||||
it("marks oversized unsafe command previews as omitted", async () => {
|
||||
@@ -355,11 +358,21 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
expect(gatewayRequestPayload().description).toBe(
|
||||
"Command: [preview truncated or unsafe content omitted]\nSession: agent:main:session-1",
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description:
|
||||
"Command: [preview truncated or unsafe content omitted]\nSession: agent:main:session-1",
|
||||
}),
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({
|
||||
commandPreviewOmitted: true,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const omittedEvent = findApprovalEvent(params, {});
|
||||
expect(omittedEvent.commandPreviewOmitted).toBe(true);
|
||||
});
|
||||
|
||||
it("marks clipped command previews even when a safe prefix remains", async () => {
|
||||
@@ -384,8 +397,14 @@ describe("Codex app-server approval bridge", () => {
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
const description = (requestPayload as { description: string }).description;
|
||||
expect(description).toContain("[preview truncated or unsafe content omitted]");
|
||||
const omittedEvent = findApprovalEvent(params, {});
|
||||
expect(omittedEvent.commandPreviewOmitted).toBe(true);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({
|
||||
commandPreviewOmitted: true,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("does not trust request-time decisions for two-phase command approvals", async () => {
|
||||
@@ -416,12 +435,15 @@ describe("Codex app-server approval bridge", () => {
|
||||
"plugin.approval.request",
|
||||
"plugin.approval.waitDecision",
|
||||
]);
|
||||
expect(
|
||||
findApprovalEvent(params, {
|
||||
status: "denied",
|
||||
approvalId: "plugin:approval-untrusted",
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({
|
||||
status: "denied",
|
||||
approvalId: "plugin:approval-untrusted",
|
||||
}),
|
||||
}),
|
||||
).toBeDefined();
|
||||
);
|
||||
});
|
||||
|
||||
it("only treats own null data-property request decisions as no-route", async () => {
|
||||
@@ -538,9 +560,12 @@ describe("Codex app-server approval bridge", () => {
|
||||
|
||||
expect(result).toEqual({ decision: "decline" });
|
||||
expect(mockCallGatewayTool).toHaveBeenCalledTimes(1);
|
||||
expect(
|
||||
findApprovalEvent(params, { status: "unavailable", reason: "needs write access" }),
|
||||
).toBeDefined();
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({ status: "unavailable", reason: "needs write access" }),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("sanitizes reason previews before forwarding approval text and events", async () => {
|
||||
@@ -563,15 +588,21 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
expect(gatewayRequestPayload().description).toBe(
|
||||
"Reason: needs write access for /tmp please\nSession: agent:main:session-1",
|
||||
);
|
||||
expect(
|
||||
findApprovalEvent(params, {
|
||||
status: "unavailable",
|
||||
reason: "needs write access for /tmp please",
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: "Reason: needs write access for /tmp please\nSession: agent:main:session-1",
|
||||
}),
|
||||
).toBeDefined();
|
||||
);
|
||||
expect(params.onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "approval",
|
||||
data: expect.objectContaining({
|
||||
status: "unavailable",
|
||||
reason: "needs write access for /tmp please",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("fails closed for unsupported native approval methods without requesting plugin approval", async () => {
|
||||
@@ -625,20 +656,28 @@ describe("Codex app-server approval bridge", () => {
|
||||
},
|
||||
scope: "turn",
|
||||
});
|
||||
expect(mockCallGatewayTool.mock.calls[0]?.[0]).toBe("plugin.approval.request");
|
||||
expect(typeof mockCallGatewayTool.mock.calls[0]?.[1]).toBe("object");
|
||||
const requestPayload = gatewayRequestPayload();
|
||||
expect(requestPayload.title).toBe("Codex app-server permission approval");
|
||||
expect(requestPayload.toolName).toBe("codex_permission_approval");
|
||||
const description = String(requestPayload.description);
|
||||
expect(description).toContain("Permissions: network, fileSystem");
|
||||
expect(gatewayCallOptions()).toEqual({ expectFinal: false });
|
||||
expect(mockCallGatewayTool).toHaveBeenCalledWith(
|
||||
"plugin.approval.request",
|
||||
expect.any(Object),
|
||||
expect.objectContaining({
|
||||
title: "Codex app-server permission approval",
|
||||
toolName: "codex_permission_approval",
|
||||
description: expect.stringContaining("Permissions: network, fileSystem"),
|
||||
}),
|
||||
{ expectFinal: false },
|
||||
);
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
const description = (requestPayload as { description: string }).description;
|
||||
expect(description).toContain("Network allowHosts: example.com, *.internal");
|
||||
expect(description).toContain("File system roots: /; writePaths: ~");
|
||||
expect(description).toContain(
|
||||
"High-risk targets: wildcard hosts, private-network wildcards, filesystem root",
|
||||
);
|
||||
expect(description).not.toContain("agent:main:session-1");
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: expect.not.stringContaining("agent:main:session-1"),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps permission detail bounded with truncated and compacted target samples", async () => {
|
||||
@@ -674,7 +713,13 @@ describe("Codex app-server approval bridge", () => {
|
||||
turnId: "turn-1",
|
||||
});
|
||||
|
||||
const description = String(gatewayRequestPayload().description);
|
||||
const [, , requestPayload] = mockCallGatewayTool.mock.calls[0] ?? [];
|
||||
expect(requestPayload).toEqual(
|
||||
expect.objectContaining({
|
||||
description: expect.any(String),
|
||||
}),
|
||||
);
|
||||
const description = (requestPayload as { description: string }).description;
|
||||
expect(description.length).toBeLessThanOrEqual(700);
|
||||
expect(description).toContain("example.com");
|
||||
expect(description).not.toContain("secret-token");
|
||||
|
||||
@@ -18,42 +18,6 @@ function resolveRuntimeForTest(params: RuntimeOptionsParams = {}) {
|
||||
return resolveCodexAppServerRuntimeOptions({ env: {}, requirementsToml: null, ...params });
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function expectFields(
|
||||
value: unknown,
|
||||
label: string,
|
||||
fields: Record<string, unknown>,
|
||||
): Record<string, unknown> {
|
||||
const record = requireRecord(value, label);
|
||||
for (const [key, expected] of Object.entries(fields)) {
|
||||
expect(record[key]).toEqual(expected);
|
||||
}
|
||||
return record;
|
||||
}
|
||||
|
||||
function expectRuntimePolicy(
|
||||
runtime: unknown,
|
||||
fields: {
|
||||
approvalPolicy: string;
|
||||
sandbox: string;
|
||||
approvalsReviewer: string;
|
||||
},
|
||||
) {
|
||||
expectFields(runtime, "runtime policy", fields);
|
||||
}
|
||||
|
||||
function expectUiHintLabel(manifest: { uiHints: Record<string, unknown> }, key: string) {
|
||||
const hint = requireRecord(manifest.uiHints[key], `${key} UI hint`);
|
||||
expect(typeof hint.label).toBe("string");
|
||||
expect((hint.label as string).length).toBeGreaterThan(0);
|
||||
}
|
||||
|
||||
describe("Codex app-server config", () => {
|
||||
it("parses typed plugin config before falling back to environment knobs", () => {
|
||||
const runtime = resolveRuntimeForTest({
|
||||
@@ -76,18 +40,20 @@ describe("Codex app-server config", () => {
|
||||
},
|
||||
});
|
||||
|
||||
expectFields(runtime, "runtime", {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "guardian_subagent",
|
||||
serviceTier: "flex",
|
||||
turnCompletionIdleTimeoutMs: 120_000,
|
||||
});
|
||||
expectFields(runtime.start, "runtime start", {
|
||||
transport: "websocket",
|
||||
url: "ws://127.0.0.1:39175",
|
||||
headers: { "X-Test": "yes" },
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "guardian_subagent",
|
||||
serviceTier: "flex",
|
||||
turnCompletionIdleTimeoutMs: 120_000,
|
||||
start: expect.objectContaining({
|
||||
transport: "websocket",
|
||||
url: "ws://127.0.0.1:39175",
|
||||
headers: { "X-Test": "yes" },
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("ignores app-server environment clearing for websocket transports", () => {
|
||||
@@ -115,9 +81,11 @@ describe("Codex app-server config", () => {
|
||||
env: {},
|
||||
});
|
||||
|
||||
expectFields(runtime.start, "runtime start", {
|
||||
clearEnv: ["OPENAI_API_KEY"],
|
||||
});
|
||||
expect(runtime.start).toEqual(
|
||||
expect.objectContaining({
|
||||
clearEnv: ["OPENAI_API_KEY"],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("normalizes legacy service tiers without discarding the rest of the config", () => {
|
||||
@@ -133,12 +101,14 @@ describe("Codex app-server config", () => {
|
||||
env: {},
|
||||
});
|
||||
|
||||
expectFields(runtime, "runtime", {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "read-only",
|
||||
approvalsReviewer: "auto_review",
|
||||
serviceTier: "priority",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "read-only",
|
||||
approvalsReviewer: "auto_review",
|
||||
serviceTier: "priority",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("passes through non-empty Codex app-server service tiers for forward compatibility", () => {
|
||||
@@ -178,15 +148,17 @@ describe("Codex app-server config", () => {
|
||||
pluginConfig: {},
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expectFields(runtime.start, "runtime start", {
|
||||
command: "codex",
|
||||
commandSource: "managed",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
start: expect.objectContaining({
|
||||
command: "codex",
|
||||
commandSource: "managed",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("defaults native Codex approvals to guardian when requirements disallow full access", () => {
|
||||
@@ -195,11 +167,13 @@ describe("Codex app-server config", () => {
|
||||
requirementsToml: 'allowed_sandbox_modes = ["read-only", "workspace-write"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("uses read-only sandbox for guardian defaults when requirements only allow read-only", () => {
|
||||
@@ -208,11 +182,13 @@ describe("Codex app-server config", () => {
|
||||
requirementsToml: 'allowed_sandbox_modes = ["read-only"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "read-only",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "read-only",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("defaults native Codex approvals to guardian when requirements disallow never approval", () => {
|
||||
@@ -221,11 +197,13 @@ describe("Codex app-server config", () => {
|
||||
requirementsToml: 'allowed_approval_policies = ["on-request"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("selects an allowed guardian approval policy when on-request is unavailable", () => {
|
||||
@@ -234,11 +212,13 @@ describe("Codex app-server config", () => {
|
||||
requirementsToml: 'allowed_approval_policies = ["on-failure"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-failure",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-failure",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps native Codex approvals unchained when requirements allow never approval", () => {
|
||||
@@ -247,11 +227,13 @@ describe("Codex app-server config", () => {
|
||||
requirementsToml: 'allowed_approval_policies = ["never"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("defaults native Codex approvals to guardian when requirements disallow user reviewer", () => {
|
||||
@@ -260,11 +242,13 @@ describe("Codex app-server config", () => {
|
||||
requirementsToml: 'allowed_approvals_reviewers = ["auto_review"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("selects an allowed reviewer when sandbox requirements force guardian defaults", () => {
|
||||
@@ -274,11 +258,13 @@ describe("Codex app-server config", () => {
|
||||
'allowed_sandbox_modes = ["read-only", "workspace-write"]\nallowed_approvals_reviewers = ["user"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("ignores quoted sandbox modes inside requirements comments", () => {
|
||||
@@ -292,11 +278,13 @@ describe("Codex app-server config", () => {
|
||||
`,
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("applies the first matching remote sandbox requirements before resolving local stdio defaults", () => {
|
||||
@@ -313,11 +301,13 @@ allowed_sandbox_modes = ["read-only", "danger-full-access"]
|
||||
`,
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("ignores non-matching remote-only sandbox requirements when resolving local stdio defaults", () => {
|
||||
@@ -330,11 +320,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
`,
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("reads local requirements policy from the configured requirements path", () => {
|
||||
@@ -350,11 +342,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
});
|
||||
|
||||
expect(readPaths).toEqual(["/custom/codex/requirements.toml"]);
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("reads local requirements policy from the Codex Windows requirements path", () => {
|
||||
@@ -370,11 +364,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
});
|
||||
|
||||
expect(readPaths).toEqual(["D:\\ManagedData\\OpenAI\\Codex\\requirements.toml"]);
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps native Codex approvals unchained when requirements allow full access", () => {
|
||||
@@ -384,11 +380,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
'allowed_sandbox_modes = ["ReadOnly", "WorkspaceWrite", "DangerFullAccess"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps native Codex approvals unchained when requirements are malformed", () => {
|
||||
@@ -397,11 +395,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
requirementsToml: "allowed_sandbox_modes = [read-only]\n",
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("does not apply local requirements policy to websocket app-server transports", () => {
|
||||
@@ -415,37 +415,41 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
requirementsToml: 'allowed_sandbox_modes = ["read-only", "workspace-write"]\n',
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps explicit yolo mode when requirements disallow full access", () => {
|
||||
const requirementsToml = 'allowed_sandbox_modes = ["read-only", "workspace-write"]\n';
|
||||
expectRuntimePolicy(
|
||||
expect(
|
||||
resolveRuntimeForTest({
|
||||
pluginConfig: { appServer: { mode: "yolo" } },
|
||||
requirementsToml,
|
||||
}),
|
||||
{
|
||||
).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
},
|
||||
}),
|
||||
);
|
||||
expectRuntimePolicy(
|
||||
expect(
|
||||
resolveRuntimeForTest({
|
||||
pluginConfig: {},
|
||||
env: { OPENCLAW_CODEX_APP_SERVER_MODE: "yolo" },
|
||||
requirementsToml,
|
||||
}),
|
||||
{
|
||||
).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
},
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -536,28 +540,28 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
});
|
||||
|
||||
it("treats configured and environment commands as explicit overrides", () => {
|
||||
expectFields(
|
||||
expect(
|
||||
resolveRuntimeForTest({
|
||||
pluginConfig: { appServer: { command: "/opt/codex/bin/codex" } },
|
||||
env: { OPENCLAW_CODEX_APP_SERVER_BIN: "/usr/local/bin/codex" },
|
||||
}).start,
|
||||
"configured start",
|
||||
{
|
||||
).toEqual(
|
||||
expect.objectContaining({
|
||||
command: "/opt/codex/bin/codex",
|
||||
commandSource: "config",
|
||||
},
|
||||
}),
|
||||
);
|
||||
|
||||
expectFields(
|
||||
expect(
|
||||
resolveRuntimeForTest({
|
||||
pluginConfig: {},
|
||||
env: { OPENCLAW_CODEX_APP_SERVER_BIN: "/usr/local/bin/codex" },
|
||||
}).start,
|
||||
"environment start",
|
||||
{
|
||||
).toEqual(
|
||||
expect.objectContaining({
|
||||
command: "/usr/local/bin/codex",
|
||||
commandSource: "env",
|
||||
},
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -583,7 +587,7 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
marketplaceName: "desktop-tools",
|
||||
});
|
||||
|
||||
expectFields(
|
||||
expect(
|
||||
resolveCodexComputerUseConfig({
|
||||
pluginConfig: {},
|
||||
env: {
|
||||
@@ -593,13 +597,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
OPENCLAW_CODEX_COMPUTER_USE_MARKETPLACE_DISCOVERY_TIMEOUT_MS: "30000",
|
||||
},
|
||||
}),
|
||||
"computer use config",
|
||||
{
|
||||
).toEqual(
|
||||
expect.objectContaining({
|
||||
enabled: true,
|
||||
autoInstall: true,
|
||||
marketplaceDiscoveryTimeoutMs: 30_000,
|
||||
marketplaceSource: "github:example/plugins",
|
||||
},
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -613,11 +617,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
env: {},
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("allows environment mode fallback to opt in to guardian-reviewed local execution", () => {
|
||||
@@ -626,11 +632,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
env: { OPENCLAW_CODEX_APP_SERVER_MODE: "guardian" },
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-request",
|
||||
sandbox: "workspace-write",
|
||||
approvalsReviewer: "auto_review",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("accepts the latest auto_review reviewer and legacy guardian_subagent alias", () => {
|
||||
@@ -654,11 +662,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
env: { OPENCLAW_CODEX_APP_SERVER_GUARDIAN: "1" },
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "never",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("lets explicit policy fields override guardian mode", () => {
|
||||
@@ -674,11 +684,13 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
env: {},
|
||||
});
|
||||
|
||||
expectRuntimePolicy(runtime, {
|
||||
approvalPolicy: "on-failure",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
});
|
||||
expect(runtime).toEqual(
|
||||
expect.objectContaining({
|
||||
approvalPolicy: "on-failure",
|
||||
sandbox: "danger-full-access",
|
||||
approvalsReviewer: "user",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("derives distinct shared-client keys for distinct auth tokens without exposing them", () => {
|
||||
@@ -779,21 +791,27 @@ allowed_sandbox_modes = ["read-only", "workspace-write"]
|
||||
|
||||
expect(manifestKeys).toEqual([...CODEX_APP_SERVER_CONFIG_KEYS].toSorted());
|
||||
for (const key of CODEX_APP_SERVER_CONFIG_KEYS) {
|
||||
expectUiHintLabel(manifest, `appServer.${key}`);
|
||||
expect(manifest.uiHints[`appServer.${key}`]).toMatchObject({
|
||||
label: expect.any(String),
|
||||
});
|
||||
}
|
||||
const computerUseManifestKeys = Object.keys(
|
||||
manifest.configSchema.properties.computerUse.properties,
|
||||
).toSorted();
|
||||
expect(computerUseManifestKeys).toEqual([...CODEX_COMPUTER_USE_CONFIG_KEYS].toSorted());
|
||||
for (const key of CODEX_COMPUTER_USE_CONFIG_KEYS) {
|
||||
expectUiHintLabel(manifest, `computerUse.${key}`);
|
||||
expect(manifest.uiHints[`computerUse.${key}`]).toMatchObject({
|
||||
label: expect.any(String),
|
||||
});
|
||||
}
|
||||
const codexPluginsProperties = manifest.configSchema.properties.codexPlugins;
|
||||
const codexPluginsManifestKeys = Object.keys(codexPluginsProperties.properties).toSorted();
|
||||
expect(codexPluginsManifestKeys).toEqual([...CODEX_PLUGINS_CONFIG_KEYS].toSorted());
|
||||
expect(codexPluginsProperties.additionalProperties).toBe(false);
|
||||
for (const key of CODEX_PLUGINS_CONFIG_KEYS) {
|
||||
expectUiHintLabel(manifest, `codexPlugins.${key}`);
|
||||
expect(manifest.uiHints[`codexPlugins.${key}`]).toMatchObject({
|
||||
label: expect.any(String),
|
||||
});
|
||||
}
|
||||
const pluginEntryProperties = (
|
||||
codexPluginsProperties.properties.plugins as {
|
||||
|
||||
@@ -117,84 +117,6 @@ function buildEmptyToolTelemetry(): CodexAppServerToolTelemetry {
|
||||
};
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): unknown[] {
|
||||
if (!Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value;
|
||||
}
|
||||
|
||||
function expectUsageFields(
|
||||
usage: unknown,
|
||||
expected: { input: number; output: number; cacheRead: number; total: number },
|
||||
) {
|
||||
const record = requireRecord(usage, "usage");
|
||||
expect(record.input).toBe(expected.input);
|
||||
expect(record.output).toBe(expected.output);
|
||||
expect(record.cacheRead).toBe(expected.cacheRead);
|
||||
expect(record.total ?? record.totalTokens).toBe(expected.total);
|
||||
}
|
||||
|
||||
function mockCallArg(mock: unknown, callIndex: number, argIndex: number, label: string) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`Expected ${label} mock calls`);
|
||||
}
|
||||
const call = calls[callIndex];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} call ${callIndex + 1}`);
|
||||
}
|
||||
return call[argIndex];
|
||||
}
|
||||
|
||||
function findAgentEvent(
|
||||
mock: unknown,
|
||||
params: { stream: string; phase?: string; itemId?: string; name?: string },
|
||||
) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error("Expected onAgentEvent mock calls");
|
||||
}
|
||||
for (const call of calls) {
|
||||
const event = requireRecord(call[0], "agent event");
|
||||
const data = requireRecord(event.data, "agent event data");
|
||||
if (
|
||||
event.stream === params.stream &&
|
||||
(!params.phase || data.phase === params.phase) &&
|
||||
(!params.itemId || data.itemId === params.itemId) &&
|
||||
(!params.name || data.name === params.name)
|
||||
) {
|
||||
return { event, data };
|
||||
}
|
||||
}
|
||||
throw new Error(`Expected agent event ${params.stream}`);
|
||||
}
|
||||
|
||||
function findPlanEventWithSteps(mock: unknown, steps: string[]) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error("Expected onAgentEvent mock calls");
|
||||
}
|
||||
for (const call of calls) {
|
||||
const event = requireRecord(call[0], "agent event");
|
||||
if (event.stream !== "plan") {
|
||||
continue;
|
||||
}
|
||||
const data = requireRecord(event.data, "plan event data");
|
||||
if (JSON.stringify(data.steps) === JSON.stringify(steps)) {
|
||||
return data;
|
||||
}
|
||||
}
|
||||
throw new Error(`Expected plan event ${steps.join(", ")}`);
|
||||
}
|
||||
|
||||
function forCurrentTurn(
|
||||
method: ProjectorNotification["method"],
|
||||
params: Record<string, unknown>,
|
||||
@@ -283,12 +205,12 @@ describe("CodexAppServerEventProjector", () => {
|
||||
expect(result.assistantTexts).toEqual(["hello"]);
|
||||
expect(result.messagesSnapshot.map((message) => message.role)).toEqual(["user", "assistant"]);
|
||||
expect(result.lastAssistant?.content).toEqual([{ type: "text", text: "hello" }]);
|
||||
expectUsageFields(result.attemptUsage, { input: 3, output: 7, cacheRead: 2, total: 12 });
|
||||
expectUsageFields(result.lastAssistant?.usage, {
|
||||
expect(result.attemptUsage).toMatchObject({ input: 3, output: 7, cacheRead: 2, total: 12 });
|
||||
expect(result.lastAssistant?.usage).toMatchObject({
|
||||
input: 3,
|
||||
output: 7,
|
||||
cacheRead: 2,
|
||||
total: 12,
|
||||
totalTokens: 12,
|
||||
});
|
||||
expect(result.replayMetadata.replaySafe).toBe(true);
|
||||
});
|
||||
@@ -314,11 +236,11 @@ describe("CodexAppServerEventProjector", () => {
|
||||
|
||||
expect(result.assistantTexts).toEqual(["done"]);
|
||||
expect(result.attemptUsage).toBeUndefined();
|
||||
expectUsageFields(result.lastAssistant?.usage, {
|
||||
expect(result.lastAssistant?.usage).toMatchObject({
|
||||
input: 0,
|
||||
output: 0,
|
||||
cacheRead: 0,
|
||||
total: 0,
|
||||
totalTokens: 0,
|
||||
});
|
||||
});
|
||||
|
||||
@@ -343,55 +265,6 @@ describe("CodexAppServerEventProjector", () => {
|
||||
expect(result.lastAssistant?.content).toEqual([{ type: "text", text: "OK from raw" }]);
|
||||
});
|
||||
|
||||
it("attaches native Codex image-generation saved paths as reply media", async () => {
|
||||
const projector = await createProjector();
|
||||
const savedPath = "/tmp/codex-home/generated_images/session-1/ig_123.png";
|
||||
|
||||
await projector.handleNotification(
|
||||
turnCompleted([
|
||||
{
|
||||
type: "imageGeneration",
|
||||
id: "ig_123",
|
||||
status: "completed",
|
||||
revisedPrompt: "A tiny blue square",
|
||||
result: "Zm9v",
|
||||
savedPath,
|
||||
},
|
||||
]),
|
||||
);
|
||||
|
||||
const result = projector.buildResult(buildEmptyToolTelemetry());
|
||||
|
||||
expect(result.assistantTexts).toStrictEqual([]);
|
||||
expect(result.toolMediaUrls).toEqual([savedPath]);
|
||||
});
|
||||
|
||||
it("does not append native Codex image-generation media after explicit media delivery", async () => {
|
||||
const projector = await createProjector();
|
||||
const savedPath = "/tmp/codex-home/generated_images/session-1/ig_123.png";
|
||||
|
||||
await projector.handleNotification(
|
||||
turnCompleted([
|
||||
{
|
||||
type: "imageGeneration",
|
||||
id: "ig_123",
|
||||
status: "completed",
|
||||
revisedPrompt: null,
|
||||
result: "Zm9v",
|
||||
savedPath,
|
||||
},
|
||||
]),
|
||||
);
|
||||
|
||||
const result = projector.buildResult({
|
||||
...buildEmptyToolTelemetry(),
|
||||
messagingToolSentMediaUrls: [savedPath],
|
||||
toolMediaUrls: [],
|
||||
});
|
||||
|
||||
expect(result.toolMediaUrls).toStrictEqual([]);
|
||||
});
|
||||
|
||||
it("does not fail a completed reply after a retryable app-server error notification", async () => {
|
||||
const projector = await createProjector();
|
||||
|
||||
@@ -535,12 +408,12 @@ describe("CodexAppServerEventProjector", () => {
|
||||
|
||||
const result = projector.buildResult(buildEmptyToolTelemetry());
|
||||
|
||||
expectUsageFields(result.attemptUsage, { input: 5, output: 9, cacheRead: 3, total: 17 });
|
||||
expectUsageFields(result.lastAssistant?.usage, {
|
||||
expect(result.attemptUsage).toMatchObject({ input: 5, output: 9, cacheRead: 3, total: 17 });
|
||||
expect(result.lastAssistant?.usage).toMatchObject({
|
||||
input: 5,
|
||||
output: 9,
|
||||
cacheRead: 3,
|
||||
total: 17,
|
||||
totalTokens: 17,
|
||||
});
|
||||
});
|
||||
|
||||
@@ -683,26 +556,30 @@ describe("CodexAppServerEventProjector", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const started = findAgentEvent(onAgentEvent, {
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "codex_app_server.guardian",
|
||||
phase: "started",
|
||||
}).data;
|
||||
expect(started.reviewId).toBe("review-1");
|
||||
expect(started.targetItemId).toBe("cmd-1");
|
||||
expect(started.status).toBe("inProgress");
|
||||
expect(started.actionType).toBe("execve");
|
||||
const completed = findAgentEvent(onAgentEvent, {
|
||||
data: expect.objectContaining({
|
||||
phase: "started",
|
||||
reviewId: "review-1",
|
||||
targetItemId: "cmd-1",
|
||||
status: "inProgress",
|
||||
actionType: "execve",
|
||||
}),
|
||||
});
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "codex_app_server.guardian",
|
||||
phase: "completed",
|
||||
}).data;
|
||||
expect(completed.reviewId).toBe("review-1");
|
||||
expect(completed.targetItemId).toBe("cmd-1");
|
||||
expect(completed.decisionSource).toBe("agent");
|
||||
expect(completed.status).toBe("approved");
|
||||
expect(completed.riskLevel).toBe("low");
|
||||
expect(completed.userAuthorization).toBe("high");
|
||||
expect(completed.rationale).toBe("Benign local probe.");
|
||||
expect(completed.actionType).toBe("execve");
|
||||
data: expect.objectContaining({
|
||||
phase: "completed",
|
||||
reviewId: "review-1",
|
||||
targetItemId: "cmd-1",
|
||||
decisionSource: "agent",
|
||||
status: "approved",
|
||||
riskLevel: "low",
|
||||
userAuthorization: "high",
|
||||
rationale: "Benign local probe.",
|
||||
actionType: "execve",
|
||||
}),
|
||||
});
|
||||
expect(
|
||||
projector.buildResult(buildEmptyToolTelemetry()).didSendDeterministicApprovalPrompt,
|
||||
).toBe(false);
|
||||
@@ -760,11 +637,17 @@ describe("CodexAppServerEventProjector", () => {
|
||||
|
||||
expect(onReasoningStream).toHaveBeenCalledWith({ text: "thinking" });
|
||||
expect(onReasoningEnd).toHaveBeenCalledTimes(1);
|
||||
expect(findPlanEventWithSteps(onAgentEvent, ["patch (in_progress)"]).steps).toEqual([
|
||||
"patch (in_progress)",
|
||||
]);
|
||||
expect(findAgentEvent(onAgentEvent, { stream: "compaction", phase: "start" }).data.itemId).toBe(
|
||||
"compact-1",
|
||||
expect(onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "plan",
|
||||
data: expect.objectContaining({ steps: ["patch (in_progress)"] }),
|
||||
}),
|
||||
);
|
||||
expect(onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "compaction",
|
||||
data: expect.objectContaining({ phase: "start", itemId: "compact-1" }),
|
||||
}),
|
||||
);
|
||||
expect(result.toolMetas).toEqual([{ toolName: "sessions_send" }]);
|
||||
expect(result.messagesSnapshot.map((message) => message.role)).toEqual([
|
||||
@@ -774,7 +657,7 @@ describe("CodexAppServerEventProjector", () => {
|
||||
]);
|
||||
expect(JSON.stringify(result.messagesSnapshot[1])).toContain("Codex reasoning");
|
||||
expect(JSON.stringify(result.messagesSnapshot[2])).toContain("Codex plan");
|
||||
expect(requireRecord(result.itemLifecycle, "item lifecycle").compactionCount).toBe(1);
|
||||
expect(result.itemLifecycle).toMatchObject({ compactionCount: 1 });
|
||||
});
|
||||
|
||||
it("synthesizes normalized tool progress for Codex-native tool items", async () => {
|
||||
@@ -816,63 +699,68 @@ describe("CodexAppServerEventProjector", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const itemStart = findAgentEvent(onAgentEvent, {
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "item",
|
||||
phase: "start",
|
||||
itemId: "cmd-1",
|
||||
}).data;
|
||||
expect(itemStart.kind).toBe("command");
|
||||
expect(itemStart.name).toBe("bash");
|
||||
expect(itemStart.suppressChannelProgress).toBe(true);
|
||||
const toolStart = findAgentEvent(onAgentEvent, {
|
||||
data: expect.objectContaining({
|
||||
phase: "start",
|
||||
kind: "command",
|
||||
name: "bash",
|
||||
itemId: "cmd-1",
|
||||
suppressChannelProgress: true,
|
||||
}),
|
||||
});
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "tool",
|
||||
phase: "start",
|
||||
itemId: "cmd-1",
|
||||
name: "bash",
|
||||
}).data;
|
||||
expect(toolStart.toolCallId).toBe("cmd-1");
|
||||
expect(toolStart.args).toEqual({ command: "pnpm test extensions/codex", cwd: "/workspace" });
|
||||
const toolResult = findAgentEvent(onAgentEvent, {
|
||||
data: expect.objectContaining({
|
||||
phase: "start",
|
||||
name: "bash",
|
||||
itemId: "cmd-1",
|
||||
toolCallId: "cmd-1",
|
||||
args: { command: "pnpm test extensions/codex", cwd: "/workspace" },
|
||||
}),
|
||||
});
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "tool",
|
||||
phase: "result",
|
||||
itemId: "cmd-1",
|
||||
name: "bash",
|
||||
}).data;
|
||||
expect(toolResult.toolCallId).toBe("cmd-1");
|
||||
expect(toolResult.status).toBe("completed");
|
||||
expect(toolResult.isError).toBe(false);
|
||||
const toolResultPayload = requireRecord(toolResult.result, "tool result payload");
|
||||
expect(toolResultPayload.exitCode).toBe(0);
|
||||
expect(toolResultPayload.durationMs).toBe(42);
|
||||
data: expect.objectContaining({
|
||||
phase: "result",
|
||||
name: "bash",
|
||||
itemId: "cmd-1",
|
||||
toolCallId: "cmd-1",
|
||||
status: "completed",
|
||||
isError: false,
|
||||
result: expect.objectContaining({ exitCode: 0, durationMs: 42 }),
|
||||
}),
|
||||
});
|
||||
const result = projector.buildResult(buildEmptyToolTelemetry());
|
||||
expect(result.messagesSnapshot.map((message) => message.role)).toEqual([
|
||||
"user",
|
||||
"assistant",
|
||||
"toolResult",
|
||||
]);
|
||||
const assistant = requireRecord(result.messagesSnapshot[1], "assistant tool call message");
|
||||
expect(assistant.role).toBe("assistant");
|
||||
const assistantContent = requireArray(assistant.content, "assistant content");
|
||||
expect(assistantContent[0]).toEqual({
|
||||
type: "toolCall",
|
||||
id: "cmd-1",
|
||||
name: "bash",
|
||||
arguments: { command: "pnpm test extensions/codex", cwd: "/workspace" },
|
||||
input: { command: "pnpm test extensions/codex", cwd: "/workspace" },
|
||||
expect(result.messagesSnapshot[1]).toMatchObject({
|
||||
role: "assistant",
|
||||
content: [
|
||||
{
|
||||
type: "toolCall",
|
||||
id: "cmd-1",
|
||||
name: "bash",
|
||||
arguments: { command: "pnpm test extensions/codex", cwd: "/workspace" },
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(result.messagesSnapshot[2]).toMatchObject({
|
||||
role: "toolResult",
|
||||
toolCallId: "cmd-1",
|
||||
toolName: "bash",
|
||||
isError: false,
|
||||
content: [
|
||||
expect.objectContaining({
|
||||
type: "toolResult",
|
||||
toolCallId: "cmd-1",
|
||||
content: "ok",
|
||||
}),
|
||||
],
|
||||
});
|
||||
const toolResultMessage = requireRecord(result.messagesSnapshot[2], "tool result message");
|
||||
expect(toolResultMessage.role).toBe("toolResult");
|
||||
expect(toolResultMessage.toolCallId).toBe("cmd-1");
|
||||
expect(toolResultMessage.toolName).toBe("bash");
|
||||
expect(toolResultMessage.isError).toBe(false);
|
||||
const toolResultContent = requireArray(toolResultMessage.content, "tool result content");
|
||||
const toolResultContentItem = requireRecord(toolResultContent[0], "tool result content item");
|
||||
expect(toolResultContentItem.type).toBe("toolResult");
|
||||
expect(toolResultContentItem.id).toBe("cmd-1");
|
||||
expect(toolResultContentItem.name).toBe("bash");
|
||||
expect(toolResultContentItem.toolName).toBe("bash");
|
||||
expect(toolResultContentItem.toolCallId).toBe("cmd-1");
|
||||
expect(toolResultContentItem.content).toBe("ok");
|
||||
});
|
||||
|
||||
it("records dynamic OpenClaw tool calls in mirrored transcript snapshots", async () => {
|
||||
@@ -899,30 +787,30 @@ describe("CodexAppServerEventProjector", () => {
|
||||
"toolResult",
|
||||
"assistant",
|
||||
]);
|
||||
const assistant = requireRecord(result.messagesSnapshot[1], "assistant tool call message");
|
||||
expect(assistant.role).toBe("assistant");
|
||||
expect(requireArray(assistant.content, "assistant content")[0]).toEqual({
|
||||
type: "toolCall",
|
||||
id: "call-browser-1",
|
||||
name: "browser",
|
||||
arguments: { action: "open", url: "http://127.0.0.1:3000" },
|
||||
input: { action: "open", url: "http://127.0.0.1:3000" },
|
||||
expect(result.messagesSnapshot[1]).toMatchObject({
|
||||
role: "assistant",
|
||||
content: [
|
||||
{
|
||||
type: "toolCall",
|
||||
id: "call-browser-1",
|
||||
name: "browser",
|
||||
arguments: { action: "open", url: "http://127.0.0.1:3000" },
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(result.messagesSnapshot[2]).toMatchObject({
|
||||
role: "toolResult",
|
||||
toolCallId: "call-browser-1",
|
||||
toolName: "browser",
|
||||
isError: false,
|
||||
content: [
|
||||
expect.objectContaining({
|
||||
type: "toolResult",
|
||||
toolCallId: "call-browser-1",
|
||||
content: "opened",
|
||||
}),
|
||||
],
|
||||
});
|
||||
const toolResultMessage = requireRecord(result.messagesSnapshot[2], "tool result message");
|
||||
expect(toolResultMessage.role).toBe("toolResult");
|
||||
expect(toolResultMessage.toolCallId).toBe("call-browser-1");
|
||||
expect(toolResultMessage.toolName).toBe("browser");
|
||||
expect(toolResultMessage.isError).toBe(false);
|
||||
const toolResultContent = requireRecord(
|
||||
requireArray(toolResultMessage.content, "tool result content")[0],
|
||||
"tool result content item",
|
||||
);
|
||||
expect(toolResultContent.type).toBe("toolResult");
|
||||
expect(toolResultContent.id).toBe("call-browser-1");
|
||||
expect(toolResultContent.name).toBe("browser");
|
||||
expect(toolResultContent.toolName).toBe("browser");
|
||||
expect(toolResultContent.toolCallId).toBe("call-browser-1");
|
||||
expect(toolResultContent.content).toBe("opened");
|
||||
});
|
||||
|
||||
it("marks declined Codex-native tool results as non-success", async () => {
|
||||
@@ -947,24 +835,28 @@ describe("CodexAppServerEventProjector", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const itemEnd = findAgentEvent(onAgentEvent, {
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "item",
|
||||
phase: "end",
|
||||
itemId: "cmd-declined",
|
||||
}).data;
|
||||
expect(itemEnd.kind).toBe("command");
|
||||
expect(itemEnd.name).toBe("bash");
|
||||
expect(itemEnd.status).toBe("blocked");
|
||||
expect(itemEnd.suppressChannelProgress).toBe(true);
|
||||
const toolResult = findAgentEvent(onAgentEvent, {
|
||||
data: expect.objectContaining({
|
||||
phase: "end",
|
||||
kind: "command",
|
||||
name: "bash",
|
||||
itemId: "cmd-declined",
|
||||
status: "blocked",
|
||||
suppressChannelProgress: true,
|
||||
}),
|
||||
});
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "tool",
|
||||
phase: "result",
|
||||
itemId: "cmd-declined",
|
||||
name: "bash",
|
||||
}).data;
|
||||
expect(toolResult.toolCallId).toBe("cmd-declined");
|
||||
expect(toolResult.status).toBe("blocked");
|
||||
expect(toolResult.isError).toBe(true);
|
||||
data: expect.objectContaining({
|
||||
phase: "result",
|
||||
name: "bash",
|
||||
itemId: "cmd-declined",
|
||||
toolCallId: "cmd-declined",
|
||||
status: "blocked",
|
||||
isError: true,
|
||||
}),
|
||||
});
|
||||
});
|
||||
|
||||
it("leaves Codex dynamic tool item progress to item/tool/call normalization", async () => {
|
||||
@@ -987,23 +879,23 @@ describe("CodexAppServerEventProjector", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const itemStart = findAgentEvent(onAgentEvent, {
|
||||
stream: "item",
|
||||
phase: "start",
|
||||
name: "message",
|
||||
}).data;
|
||||
expect(itemStart.kind).toBe("tool");
|
||||
expect(itemStart.suppressChannelProgress).toBe(true);
|
||||
const calls = (onAgentEvent as { mock: { calls: unknown[][] } }).mock.calls;
|
||||
const toolStart = calls.some((call) => {
|
||||
const event = requireRecord(call[0], "agent event");
|
||||
if (event.stream !== "tool") {
|
||||
return false;
|
||||
}
|
||||
const data = requireRecord(event.data, "agent event data");
|
||||
return data.phase === "start" && data.name === "message";
|
||||
});
|
||||
expect(toolStart).toBe(false);
|
||||
expect(onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "item",
|
||||
data: expect.objectContaining({
|
||||
phase: "start",
|
||||
kind: "tool",
|
||||
name: "message",
|
||||
suppressChannelProgress: true,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(onAgentEvent).not.toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "tool",
|
||||
data: expect.objectContaining({ phase: "start", name: "message" }),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("emits verbose tool summaries through onToolResult", async () => {
|
||||
@@ -1255,10 +1147,12 @@ describe("CodexAppServerEventProjector", () => {
|
||||
|
||||
const result = projector.buildResult(buildEmptyToolTelemetry());
|
||||
|
||||
expect(findAgentEvent(onAgentEvent, { stream: "plan" }).data.steps).toEqual([
|
||||
"step one",
|
||||
"step two",
|
||||
]);
|
||||
expect(onAgentEvent).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
stream: "plan",
|
||||
data: expect.objectContaining({ steps: ["step one", "step two"] }),
|
||||
}),
|
||||
);
|
||||
expect(result.assistantTexts).toEqual(["final answer"]);
|
||||
expect(JSON.stringify(result.messagesSnapshot)).toContain("Codex plan");
|
||||
});
|
||||
@@ -1279,33 +1173,28 @@ describe("CodexAppServerEventProjector", () => {
|
||||
);
|
||||
expect(openSpy).not.toHaveBeenCalled();
|
||||
|
||||
const beforePayload = requireRecord(
|
||||
mockCallArg(beforeCompaction, 0, 0, "beforeCompaction"),
|
||||
"before payload",
|
||||
expect(beforeCompaction).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
messageCount: 1,
|
||||
sessionFile: expect.stringContaining("session.jsonl"),
|
||||
messages: [expect.objectContaining({ role: "assistant" })],
|
||||
}),
|
||||
expect.objectContaining({
|
||||
runId: "run-1",
|
||||
sessionId: "session-1",
|
||||
}),
|
||||
);
|
||||
expect(beforePayload.messageCount).toBe(1);
|
||||
expect(String(beforePayload.sessionFile)).toContain("session.jsonl");
|
||||
const beforeMessages = requireArray(beforePayload.messages, "before messages");
|
||||
expect(requireRecord(beforeMessages[0], "before message").role).toBe("assistant");
|
||||
const beforeContext = requireRecord(
|
||||
mockCallArg(beforeCompaction, 0, 1, "beforeCompaction"),
|
||||
"before context",
|
||||
expect(afterCompaction).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
messageCount: 1,
|
||||
compactedCount: -1,
|
||||
sessionFile: expect.stringContaining("session.jsonl"),
|
||||
}),
|
||||
expect.objectContaining({
|
||||
runId: "run-1",
|
||||
sessionId: "session-1",
|
||||
}),
|
||||
);
|
||||
expect(beforeContext.runId).toBe("run-1");
|
||||
expect(beforeContext.sessionId).toBe("session-1");
|
||||
const afterPayload = requireRecord(
|
||||
mockCallArg(afterCompaction, 0, 0, "afterCompaction"),
|
||||
"after payload",
|
||||
);
|
||||
expect(afterPayload.messageCount).toBe(1);
|
||||
expect(afterPayload.compactedCount).toBe(-1);
|
||||
expect(String(afterPayload.sessionFile)).toContain("session.jsonl");
|
||||
const afterContext = requireRecord(
|
||||
mockCallArg(afterCompaction, 0, 1, "afterCompaction"),
|
||||
"after context",
|
||||
);
|
||||
expect(afterContext.runId).toBe("run-1");
|
||||
expect(afterContext.sessionId).toBe("session-1");
|
||||
});
|
||||
|
||||
it("projects codex hook started and completed notifications into agent events", async () => {
|
||||
@@ -1347,24 +1236,28 @@ describe("CodexAppServerEventProjector", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const started = findAgentEvent(onAgentEvent, {
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "codex_app_server.hook",
|
||||
phase: "started",
|
||||
}).data;
|
||||
expect(started.threadId).toBe(THREAD_ID);
|
||||
expect(started.turnId).toBe(TURN_ID);
|
||||
expect(started.hookRunId).toBe("hook-1");
|
||||
expect(started.eventName).toBe("preToolUse");
|
||||
expect(started.status).toBe("running");
|
||||
const completed = findAgentEvent(onAgentEvent, {
|
||||
data: expect.objectContaining({
|
||||
phase: "started",
|
||||
threadId: THREAD_ID,
|
||||
turnId: TURN_ID,
|
||||
hookRunId: "hook-1",
|
||||
eventName: "preToolUse",
|
||||
status: "running",
|
||||
}),
|
||||
});
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "codex_app_server.hook",
|
||||
phase: "completed",
|
||||
}).data;
|
||||
expect(completed.hookRunId).toBe("hook-1");
|
||||
expect(completed.status).toBe("blocked");
|
||||
expect(completed.statusMessage).toBe("blocked by hook");
|
||||
expect(completed.durationMs).toBe(42);
|
||||
expect(completed.entries).toEqual([{ kind: "stderr", text: "blocked" }]);
|
||||
data: expect.objectContaining({
|
||||
phase: "completed",
|
||||
hookRunId: "hook-1",
|
||||
status: "blocked",
|
||||
statusMessage: "blocked by hook",
|
||||
durationMs: 42,
|
||||
entries: [{ kind: "stderr", text: "blocked" }],
|
||||
}),
|
||||
});
|
||||
});
|
||||
|
||||
it("projects thread-scoped codex hook notifications that omit a turn id", async () => {
|
||||
@@ -1392,14 +1285,16 @@ describe("CodexAppServerEventProjector", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const started = findAgentEvent(onAgentEvent, {
|
||||
expect(onAgentEvent).toHaveBeenCalledWith({
|
||||
stream: "codex_app_server.hook",
|
||||
phase: "started",
|
||||
}).data;
|
||||
expect(started.threadId).toBe(THREAD_ID);
|
||||
expect(started.turnId).toBeNull();
|
||||
expect(started.hookRunId).toBe("hook-thread-1");
|
||||
expect(started.eventName).toBe("sessionStart");
|
||||
expect(started.scope).toBe("thread");
|
||||
data: expect.objectContaining({
|
||||
phase: "started",
|
||||
threadId: THREAD_ID,
|
||||
turnId: null,
|
||||
hookRunId: "hook-thread-1",
|
||||
eventName: "sessionStart",
|
||||
scope: "thread",
|
||||
}),
|
||||
});
|
||||
});
|
||||
});
|
||||
|
||||
@@ -115,7 +115,6 @@ export class CodexAppServerEventProjector {
|
||||
private readonly toolTranscriptMessages: AgentMessage[] = [];
|
||||
private readonly toolTranscriptCallIds = new Set<string>();
|
||||
private readonly toolTranscriptResultIds = new Set<string>();
|
||||
private readonly nativeGeneratedMediaUrls = new Set<string>();
|
||||
private assistantStarted = false;
|
||||
private reasoningStarted = false;
|
||||
private reasoningEnded = false;
|
||||
@@ -295,7 +294,7 @@ export class CodexAppServerEventProjector {
|
||||
messagingToolSentMediaUrls: toolTelemetry.messagingToolSentMediaUrls,
|
||||
messagingToolSentTargets: toolTelemetry.messagingToolSentTargets,
|
||||
heartbeatToolResponse: toolTelemetry.heartbeatToolResponse,
|
||||
toolMediaUrls: this.buildToolMediaUrls(toolTelemetry),
|
||||
toolMediaUrls: toolTelemetry.toolMediaUrls,
|
||||
toolAudioAsVoice: toolTelemetry.toolAudioAsVoice,
|
||||
successfulCronAdds: toolTelemetry.successfulCronAdds,
|
||||
cloudCodeAssistFormatError: false,
|
||||
@@ -463,7 +462,6 @@ export class CodexAppServerEventProjector {
|
||||
this.rememberAssistantItem(item.id);
|
||||
this.assistantTextByItem.set(item.id, item.text);
|
||||
}
|
||||
this.recordNativeGeneratedMedia(item);
|
||||
if (item?.type === "plan" && typeof item.text === "string" && item.text) {
|
||||
this.planTextByItem.set(item.id, item.text);
|
||||
this.emitPlanUpdate({ explanation: undefined, steps: splitPlanText(item.text) });
|
||||
@@ -599,7 +597,6 @@ export class CodexAppServerEventProjector {
|
||||
this.rememberAssistantItem(item.id);
|
||||
this.assistantTextByItem.set(item.id, item.text);
|
||||
}
|
||||
this.recordNativeGeneratedMedia(item);
|
||||
if (item.type === "plan" && typeof item.text === "string" && item.text) {
|
||||
this.planTextByItem.set(item.id, item.text);
|
||||
this.emitPlanUpdate({ explanation: undefined, steps: splitPlanText(item.text) });
|
||||
@@ -675,28 +672,6 @@ export class CodexAppServerEventProjector {
|
||||
this.assistantTextByItem.set(itemId, text);
|
||||
}
|
||||
|
||||
private recordNativeGeneratedMedia(item: CodexThreadItem | undefined): void {
|
||||
if (item?.type !== "imageGeneration") {
|
||||
return;
|
||||
}
|
||||
const savedPath = readItemString(item, "savedPath")?.trim();
|
||||
if (savedPath) {
|
||||
this.nativeGeneratedMediaUrls.add(savedPath);
|
||||
}
|
||||
}
|
||||
|
||||
private buildToolMediaUrls(toolTelemetry: CodexAppServerToolTelemetry): string[] | undefined {
|
||||
const mediaUrls = new Set(
|
||||
toolTelemetry.toolMediaUrls?.map((url) => url.trim()).filter(Boolean) ?? [],
|
||||
);
|
||||
if ((toolTelemetry.messagingToolSentMediaUrls?.length ?? 0) === 0) {
|
||||
for (const mediaUrl of this.nativeGeneratedMediaUrls) {
|
||||
mediaUrls.add(mediaUrl);
|
||||
}
|
||||
}
|
||||
return mediaUrls.size > 0 ? [...mediaUrls] : toolTelemetry.toolMediaUrls;
|
||||
}
|
||||
|
||||
private async maybeEndReasoning(): Promise<void> {
|
||||
if (!this.reasoningStarted || this.reasoningEnded) {
|
||||
return;
|
||||
@@ -1041,7 +1016,6 @@ export class CodexAppServerEventProjector {
|
||||
}
|
||||
|
||||
private createToolCallMessage(params: ToolTranscriptCallInput): AgentMessage {
|
||||
const args = normalizeToolTranscriptArguments(params.arguments);
|
||||
return {
|
||||
role: "assistant",
|
||||
content: [
|
||||
@@ -1049,8 +1023,7 @@ export class CodexAppServerEventProjector {
|
||||
type: "toolCall",
|
||||
id: params.id,
|
||||
name: params.name,
|
||||
arguments: args,
|
||||
input: args,
|
||||
arguments: normalizeToolTranscriptArguments(params.arguments),
|
||||
},
|
||||
],
|
||||
api: this.params.model.api ?? "openai-codex-responses",
|
||||
@@ -1072,9 +1045,6 @@ export class CodexAppServerEventProjector {
|
||||
content: [
|
||||
{
|
||||
type: "toolResult",
|
||||
id: params.id,
|
||||
name: params.name,
|
||||
toolName: params.name,
|
||||
toolCallId: params.id,
|
||||
toolUseId: params.id,
|
||||
tool_use_id: params.id,
|
||||
|
||||
@@ -28,10 +28,7 @@ import {
|
||||
resolveCodexAppServerHomeDir,
|
||||
} from "./auth-bridge.js";
|
||||
import { readCodexPluginConfig, resolveCodexAppServerRuntimeOptions } from "./config.js";
|
||||
import {
|
||||
CODEX_OPENCLAW_DYNAMIC_TOOL_NAMESPACE,
|
||||
createCodexDynamicToolBridge,
|
||||
} from "./dynamic-tools.js";
|
||||
import { CODEX_OPENCLAW_DYNAMIC_TOOL_NAMESPACE } from "./dynamic-tools.js";
|
||||
import * as elicitationBridge from "./elicitation-bridge.js";
|
||||
import type { CodexServerNotification } from "./protocol.js";
|
||||
import { rememberCodexRateLimits, resetCodexRateLimitCacheForTests } from "./rate-limit-cache.js";
|
||||
@@ -636,44 +633,6 @@ describe("runCodexAppServerAttempt", () => {
|
||||
}
|
||||
});
|
||||
|
||||
it("does not expose OpenClaw Tool Search controls through Codex dynamic tools", async () => {
|
||||
const sessionFile = path.join(tempDir, "session.jsonl");
|
||||
const workspaceDir = path.join(tempDir, "workspace");
|
||||
const params = createParams(sessionFile, workspaceDir);
|
||||
params.disableTools = false;
|
||||
params.config = {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
};
|
||||
const sandboxSessionKey = params.sessionKey;
|
||||
if (!sandboxSessionKey) {
|
||||
throw new Error("createParams must provide a sessionKey for Codex dynamic tool tests.");
|
||||
}
|
||||
|
||||
const tools = await __testing.buildDynamicTools({
|
||||
params,
|
||||
resolvedWorkspace: workspaceDir,
|
||||
effectiveWorkspace: workspaceDir,
|
||||
sandboxSessionKey,
|
||||
sandbox: null as never,
|
||||
runAbortController: new AbortController(),
|
||||
sessionAgentId: "main",
|
||||
pluginConfig: {},
|
||||
onYieldDetected: () => undefined,
|
||||
});
|
||||
const bridge = createCodexDynamicToolBridge({
|
||||
tools,
|
||||
signal: new AbortController().signal,
|
||||
loading: "searchable",
|
||||
});
|
||||
const dynamicToolNames = bridge.specs.map((tool) => tool.name);
|
||||
|
||||
expect(dynamicToolNames).not.toEqual(
|
||||
expect.arrayContaining(["tool_search_code", "tool_search", "tool_describe", "tool_call"]),
|
||||
);
|
||||
});
|
||||
|
||||
it("normalizes Codex dynamic toolsAllow entries before filtering", () => {
|
||||
const tools = ["exec", "apply_patch", "read", "message"].map((name) => ({ name }));
|
||||
|
||||
@@ -734,89 +693,6 @@ describe("runCodexAppServerAttempt", () => {
|
||||
expect(heartbeat?.deferLoading).toBe(true);
|
||||
});
|
||||
|
||||
it("keeps searchable Codex dynamic tools canonical in mirrored transcript snapshots", async () => {
|
||||
__testing.setOpenClawCodingToolsFactoryForTests(() => [
|
||||
createRuntimeDynamicTool("wiki_status"),
|
||||
]);
|
||||
const harness = createStartedThreadHarness();
|
||||
const params = createParams(
|
||||
path.join(tempDir, "session.jsonl"),
|
||||
path.join(tempDir, "workspace"),
|
||||
);
|
||||
params.disableTools = false;
|
||||
params.runtimePlan = createCodexRuntimePlanFixture();
|
||||
params.toolsAllow = ["wiki_status"];
|
||||
|
||||
const run = runCodexAppServerAttempt(params, {
|
||||
pluginConfig: {
|
||||
codexDynamicToolsLoading: "searchable",
|
||||
appServer: { mode: "yolo" },
|
||||
},
|
||||
});
|
||||
await harness.waitForMethod("turn/start", 120_000);
|
||||
|
||||
const toolResult = (await harness.handleServerRequest({
|
||||
id: "request-tool-wiki-status",
|
||||
method: "item/tool/call",
|
||||
params: {
|
||||
threadId: "thread-1",
|
||||
turnId: "turn-1",
|
||||
callId: "call-wiki-status-1",
|
||||
namespace: CODEX_OPENCLAW_DYNAMIC_TOOL_NAMESPACE,
|
||||
tool: "wiki_status",
|
||||
arguments: { topic: "README.md" },
|
||||
},
|
||||
})) as {
|
||||
contentItems?: Array<{ text?: string; type?: string }>;
|
||||
success?: boolean;
|
||||
};
|
||||
expect(toolResult).toEqual({
|
||||
success: true,
|
||||
contentItems: [{ type: "inputText", text: "wiki_status done" }],
|
||||
});
|
||||
|
||||
await harness.completeTurn({ threadId: "thread-1", turnId: "turn-1" });
|
||||
const result = await run;
|
||||
|
||||
expect(result.messagesSnapshot.map((message) => message.role)).toEqual([
|
||||
"user",
|
||||
"assistant",
|
||||
"toolResult",
|
||||
]);
|
||||
expect(result.messagesSnapshot[1]).toMatchObject({
|
||||
role: "assistant",
|
||||
content: [
|
||||
{
|
||||
type: "toolCall",
|
||||
id: "call-wiki-status-1",
|
||||
name: "wiki_status",
|
||||
arguments: { topic: "README.md" },
|
||||
input: { topic: "README.md" },
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(result.messagesSnapshot[2]).toMatchObject({
|
||||
role: "toolResult",
|
||||
toolCallId: "call-wiki-status-1",
|
||||
toolName: "wiki_status",
|
||||
isError: false,
|
||||
content: [
|
||||
expect.objectContaining({
|
||||
type: "toolResult",
|
||||
id: "call-wiki-status-1",
|
||||
name: "wiki_status",
|
||||
toolName: "wiki_status",
|
||||
toolCallId: "call-wiki-status-1",
|
||||
toolUseId: "call-wiki-status-1",
|
||||
tool_use_id: "call-wiki-status-1",
|
||||
content: "wiki_status done",
|
||||
}),
|
||||
],
|
||||
});
|
||||
expect(JSON.stringify(result.messagesSnapshot)).not.toContain("tool_search");
|
||||
expect(JSON.stringify(result.messagesSnapshot)).not.toContain("function_call_output");
|
||||
});
|
||||
|
||||
it("passes the live run session key to Codex dynamic tools when sandbox policy uses another key", () => {
|
||||
const workspaceDir = path.join(tempDir, "workspace");
|
||||
const params = createParams(path.join(tempDir, "session.jsonl"), workspaceDir);
|
||||
@@ -2673,43 +2549,6 @@ describe("runCodexAppServerAttempt", () => {
|
||||
expect(result.timedOut).toBe(false);
|
||||
});
|
||||
|
||||
it("surfaces Codex-native image generation saved paths as reply media", async () => {
|
||||
const harness = createStartedThreadHarness();
|
||||
const params = createParams(
|
||||
path.join(tempDir, "session.jsonl"),
|
||||
path.join(tempDir, "workspace"),
|
||||
);
|
||||
|
||||
const run = runCodexAppServerAttempt(params);
|
||||
await harness.waitForMethod("turn/start");
|
||||
await harness.notify({
|
||||
method: "turn/completed",
|
||||
params: {
|
||||
threadId: "thread-1",
|
||||
turnId: "turn-1",
|
||||
turn: {
|
||||
id: "turn-1",
|
||||
status: "completed",
|
||||
items: [
|
||||
{
|
||||
type: "imageGeneration",
|
||||
id: "ig_123",
|
||||
status: "completed",
|
||||
revisedPrompt: "A tiny blue square",
|
||||
result: "Zm9v",
|
||||
savedPath: "/tmp/codex-home/generated_images/session-1/ig_123.png",
|
||||
},
|
||||
],
|
||||
},
|
||||
},
|
||||
});
|
||||
|
||||
await expect(run).resolves.toMatchObject({
|
||||
assistantTexts: [],
|
||||
toolMediaUrls: ["/tmp/codex-home/generated_images/session-1/ig_123.png"],
|
||||
});
|
||||
});
|
||||
|
||||
it("does not complete on unscoped turn/completed notifications", async () => {
|
||||
const harness = createStartedThreadHarness();
|
||||
const run = runCodexAppServerAttempt(
|
||||
|
||||
@@ -82,37 +82,6 @@ function createTestThreadBindingManager(
|
||||
});
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function expectFields(
|
||||
value: unknown,
|
||||
label: string,
|
||||
fields: Record<string, unknown>,
|
||||
): Record<string, unknown> {
|
||||
const record = requireRecord(value, label);
|
||||
for (const [key, expected] of Object.entries(fields)) {
|
||||
expect(record[key]).toEqual(expected);
|
||||
}
|
||||
return record;
|
||||
}
|
||||
|
||||
function mockCallArg(mock: unknown, callIndex: number, argIndex: number, label: string) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`Expected ${label} mock calls`);
|
||||
}
|
||||
const call = calls[callIndex];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} call ${callIndex + 1}`);
|
||||
}
|
||||
return call[argIndex];
|
||||
}
|
||||
|
||||
describe("thread binding lifecycle", () => {
|
||||
beforeEach(() => {
|
||||
__testing.resetThreadBindingsForTests();
|
||||
@@ -319,7 +288,7 @@ describe("thread binding lifecycle", () => {
|
||||
webhookToken: "tok-1",
|
||||
introText: "intro",
|
||||
});
|
||||
expectFields(binding, "binding", {
|
||||
expect(binding).toMatchObject({
|
||||
threadId: "thread-1",
|
||||
targetSessionKey: "agent:main:subagent:child",
|
||||
});
|
||||
@@ -361,7 +330,7 @@ describe("thread binding lifecycle", () => {
|
||||
webhookId: "wh-1",
|
||||
webhookToken: "tok-1",
|
||||
});
|
||||
expectFields(binding, "binding", {
|
||||
expect(binding).toMatchObject({
|
||||
threadId: "thread-1",
|
||||
targetSessionKey: "agent:main:subagent:child",
|
||||
});
|
||||
@@ -390,7 +359,7 @@ describe("thread binding lifecycle", () => {
|
||||
await vi.advanceTimersByTimeAsync(120_000);
|
||||
await __testing.runThreadBindingSweepForAccount("default");
|
||||
|
||||
expectFields(requireBinding(manager, "thread-1"), "thread binding", {
|
||||
expect(requireBinding(manager, "thread-1")).toMatchObject({
|
||||
threadId: "thread-1",
|
||||
targetSessionKey: "agent:main:subagent:child",
|
||||
webhookId: "wh-1",
|
||||
@@ -552,11 +521,11 @@ describe("thread binding lifecycle", () => {
|
||||
webhookToken: "tok-1",
|
||||
});
|
||||
|
||||
expectFields(rebound, "rebound binding", {
|
||||
expect(rebound).toMatchObject({
|
||||
idleTimeoutMs: 2 * 60 * 60 * 1000,
|
||||
maxAgeMs: 3 * 60 * 60 * 1000,
|
||||
});
|
||||
expectFields(requireBinding(manager, "thread-1"), "thread binding", {
|
||||
expect(requireBinding(manager, "thread-1")).toMatchObject({
|
||||
idleTimeoutMs: 2 * 60 * 60 * 1000,
|
||||
maxAgeMs: 3 * 60 * 60 * 1000,
|
||||
});
|
||||
@@ -597,7 +566,7 @@ describe("thread binding lifecycle", () => {
|
||||
await vi.advanceTimersByTimeAsync(240_000);
|
||||
await __testing.runThreadBindingSweepForAccount("default");
|
||||
|
||||
expectFields(requireBinding(manager, "thread-1"), "thread binding", {
|
||||
expect(requireBinding(manager, "thread-1")).toMatchObject({
|
||||
threadId: "thread-1",
|
||||
targetSessionKey: "agent:main:subagent:child",
|
||||
idleTimeoutMs: 0,
|
||||
@@ -661,7 +630,7 @@ describe("thread binding lifecycle", () => {
|
||||
await vi.advanceTimersByTimeAsync(120_000);
|
||||
await __testing.runThreadBindingSweepForAccount("default");
|
||||
|
||||
expectFields(requireBinding(manager, "thread-2"), "thread binding", {
|
||||
expect(requireBinding(manager, "thread-2")).toMatchObject({
|
||||
threadId: "thread-2",
|
||||
targetSessionKey: "agent:main:subagent:second",
|
||||
});
|
||||
@@ -693,7 +662,7 @@ describe("thread binding lifecycle", () => {
|
||||
|
||||
vi.setSystemTime(new Date("2026-02-20T00:00:30.000Z"));
|
||||
const touched = manager.touchThread({ threadId: "thread-1", persist: false });
|
||||
expectFields(touched, "touched binding", {
|
||||
expect(touched).toMatchObject({
|
||||
threadId: "thread-1",
|
||||
lastActivityAt: new Date("2026-02-20T00:00:30.000Z").getTime(),
|
||||
});
|
||||
@@ -786,7 +755,7 @@ describe("thread binding lifecycle", () => {
|
||||
targetSessionKey: "agent:main:subagent:child-1",
|
||||
agentId: "main",
|
||||
});
|
||||
expectFields(first, "first binding", {
|
||||
expect(first).toMatchObject({
|
||||
threadId: "thread-1",
|
||||
targetSessionKey: "agent:main:subagent:child-1",
|
||||
});
|
||||
@@ -804,7 +773,7 @@ describe("thread binding lifecycle", () => {
|
||||
targetSessionKey: "agent:main:subagent:child-2",
|
||||
agentId: "main",
|
||||
});
|
||||
expectFields(second, "second binding", {
|
||||
expect(second).toMatchObject({
|
||||
webhookId: "wh-created",
|
||||
webhookToken: "tok-created",
|
||||
});
|
||||
@@ -840,25 +809,15 @@ describe("thread binding lifecycle", () => {
|
||||
agentId: "main",
|
||||
});
|
||||
|
||||
expectFields(childBinding, "child binding", {
|
||||
expect(childBinding).toMatchObject({
|
||||
threadId: "thread-created-2",
|
||||
targetSessionKey: "agent:main:subagent:child-2",
|
||||
});
|
||||
expect(hoisted.createThreadDiscord).toHaveBeenCalledTimes(1);
|
||||
expect(mockCallArg(hoisted.createThreadDiscord, 0, 0, "createThreadDiscord")).toBe("parent-1");
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 1, "createThreadDiscord"),
|
||||
"thread options",
|
||||
{
|
||||
autoArchiveMinutes: 60,
|
||||
},
|
||||
);
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 2, "createThreadDiscord"),
|
||||
"thread context",
|
||||
{
|
||||
accountId: "default",
|
||||
},
|
||||
expect(hoisted.createThreadDiscord).toHaveBeenCalledWith(
|
||||
"parent-1",
|
||||
expect.objectContaining({ autoArchiveMinutes: 60 }),
|
||||
expect.objectContaining({ accountId: "default" }),
|
||||
);
|
||||
expect(manager.getByThreadId("thread-1")?.targetSessionKey).toBe("agent:main:subagent:parent");
|
||||
expect(manager.getByThreadId("thread-created-2")?.targetSessionKey).toBe(
|
||||
@@ -893,22 +852,12 @@ describe("thread binding lifecycle", () => {
|
||||
agentId: "main",
|
||||
});
|
||||
|
||||
expectFields(childBinding, "child binding", { channelId: "parent-1" });
|
||||
expect(childBinding).toMatchObject({ channelId: "parent-1" });
|
||||
expect(hoisted.restGet).toHaveBeenCalledTimes(1);
|
||||
expect(mockCallArg(hoisted.createThreadDiscord, 0, 0, "createThreadDiscord")).toBe("parent-1");
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 1, "createThreadDiscord"),
|
||||
"thread options",
|
||||
{
|
||||
autoArchiveMinutes: 60,
|
||||
},
|
||||
);
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 2, "createThreadDiscord"),
|
||||
"thread context",
|
||||
{
|
||||
accountId: "default",
|
||||
},
|
||||
expect(hoisted.createThreadDiscord).toHaveBeenCalledWith(
|
||||
"parent-1",
|
||||
expect.objectContaining({ autoArchiveMinutes: 60 }),
|
||||
expect.objectContaining({ accountId: "default" }),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -945,14 +894,14 @@ describe("thread binding lifecycle", () => {
|
||||
agentId: "main",
|
||||
});
|
||||
|
||||
expectFields(childBinding, "child binding", {
|
||||
expect(childBinding).toMatchObject({
|
||||
threadId: "thread-created-runtime",
|
||||
targetSessionKey: "agent:main:subagent:child-runtime",
|
||||
});
|
||||
const firstClientArgs = hoisted.createDiscordRestClient.mock.calls[0]?.[0] as
|
||||
| { accountId?: string; token?: string }
|
||||
| undefined;
|
||||
expectFields(firstClientArgs, "first client args", {
|
||||
expect(firstClientArgs).toMatchObject({
|
||||
accountId: "runtime",
|
||||
token: "runtime-token",
|
||||
});
|
||||
@@ -998,7 +947,7 @@ describe("thread binding lifecycle", () => {
|
||||
agentId: "main",
|
||||
});
|
||||
|
||||
expectFields(bound, "bound thread", {
|
||||
expect(bound).toMatchObject({
|
||||
threadId: "thread-created-runtime-cfg",
|
||||
targetSessionKey: "agent:main:subagent:runtime-cfg",
|
||||
});
|
||||
@@ -1058,27 +1007,14 @@ describe("thread binding lifecycle", () => {
|
||||
agentId: "main",
|
||||
});
|
||||
|
||||
expectFields(bound, "bound thread", {
|
||||
expect(bound).toMatchObject({
|
||||
threadId: "thread-created-token-refresh",
|
||||
targetSessionKey: "agent:main:subagent:token-refresh",
|
||||
});
|
||||
expect(mockCallArg(hoisted.createThreadDiscord, 0, 0, "createThreadDiscord")).toBe(
|
||||
expect(hoisted.createThreadDiscord).toHaveBeenCalledWith(
|
||||
"parent-runtime",
|
||||
);
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 1, "createThreadDiscord"),
|
||||
"thread options",
|
||||
{
|
||||
autoArchiveMinutes: 60,
|
||||
},
|
||||
);
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 2, "createThreadDiscord"),
|
||||
"thread context",
|
||||
{
|
||||
accountId: "runtime",
|
||||
token: "token-new",
|
||||
},
|
||||
expect.objectContaining({ autoArchiveMinutes: 60 }),
|
||||
expect.objectContaining({ accountId: "runtime", token: "token-new" }),
|
||||
);
|
||||
const usedTokenNew = hoisted.createDiscordRestClient.mock.calls.some(
|
||||
(call) => (call?.[0] as { token?: string } | undefined)?.token === "token-new",
|
||||
@@ -1116,31 +1052,17 @@ describe("thread binding lifecycle", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const boundConversation = requireRecord(
|
||||
requireRecord(bound, "bound session").conversation,
|
||||
"bound conversation",
|
||||
);
|
||||
expectFields(boundConversation, "bound conversation", {
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "thread-created-parent-normalized",
|
||||
});
|
||||
expect(mockCallArg(hoisted.createThreadDiscord, 0, 0, "createThreadDiscord")).toBe(
|
||||
"1491611525914558667",
|
||||
);
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 1, "createThreadDiscord"),
|
||||
"thread options",
|
||||
{
|
||||
autoArchiveMinutes: 60,
|
||||
},
|
||||
);
|
||||
expectFields(
|
||||
mockCallArg(hoisted.createThreadDiscord, 0, 2, "createThreadDiscord"),
|
||||
"thread context",
|
||||
{
|
||||
expect(bound).toMatchObject({
|
||||
conversation: {
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "thread-created-parent-normalized",
|
||||
},
|
||||
});
|
||||
expect(hoisted.createThreadDiscord).toHaveBeenCalledWith(
|
||||
"1491611525914558667",
|
||||
expect.objectContaining({ autoArchiveMinutes: 60 }),
|
||||
expect.objectContaining({ accountId: "default" }),
|
||||
);
|
||||
expect(hoisted.restGet).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -1172,26 +1094,22 @@ describe("thread binding lifecycle", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const boundConversation = requireRecord(
|
||||
requireRecord(bound, "bound session").conversation,
|
||||
"bound conversation",
|
||||
);
|
||||
expectFields(boundConversation, "bound conversation", {
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "channel:1491611525914558667",
|
||||
expect(bound).toMatchObject({
|
||||
conversation: {
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "channel:1491611525914558667",
|
||||
},
|
||||
});
|
||||
expectFields(
|
||||
expect(
|
||||
service.resolveByConversation({
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "channel:1491611525914558667",
|
||||
}),
|
||||
"resolved binding",
|
||||
{
|
||||
targetSessionKey: "agent:codex:acp:current-channel",
|
||||
},
|
||||
);
|
||||
).toMatchObject({
|
||||
targetSessionKey: "agent:codex:acp:current-channel",
|
||||
});
|
||||
expect(
|
||||
service.resolveByConversation({
|
||||
channel: "discord",
|
||||
@@ -1231,27 +1149,25 @@ describe("thread binding lifecycle", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const boundConversation = requireRecord(
|
||||
requireRecord(bound, "bound session").conversation,
|
||||
"bound conversation",
|
||||
);
|
||||
expectFields(boundConversation, "bound conversation", {
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "user:1177378744822943744",
|
||||
parentConversationId: "user:1177378744822943744",
|
||||
expect(bound).toMatchObject({
|
||||
conversation: {
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "user:1177378744822943744",
|
||||
parentConversationId: "user:1177378744822943744",
|
||||
},
|
||||
});
|
||||
const resolved = requireRecord(
|
||||
expect(
|
||||
getSessionBindingService().resolveByConversation({
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "user:1177378744822943744",
|
||||
}),
|
||||
"resolved binding",
|
||||
);
|
||||
expect(requireRecord(resolved.conversation, "resolved conversation").conversationId).toBe(
|
||||
"user:1177378744822943744",
|
||||
);
|
||||
).toMatchObject({
|
||||
conversation: {
|
||||
conversationId: "user:1177378744822943744",
|
||||
},
|
||||
});
|
||||
expect(hoisted.restGet).not.toHaveBeenCalled();
|
||||
expect(hoisted.restPost).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -1297,21 +1213,21 @@ describe("thread binding lifecycle", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const resolved = requireRecord(
|
||||
expect(
|
||||
getSessionBindingService().resolveByConversation({
|
||||
channel: "discord",
|
||||
accountId: "default",
|
||||
conversationId: "user:1177378744822943744",
|
||||
}),
|
||||
"resolved binding",
|
||||
);
|
||||
expectFields(requireRecord(resolved.metadata, "resolved metadata"), "resolved metadata", {
|
||||
pluginBindingOwner: "plugin",
|
||||
pluginId: "openclaw-codex-app-server",
|
||||
pluginRoot: "/Users/huntharo/github/openclaw-app-server",
|
||||
agentId: "codex",
|
||||
boundBy: "system",
|
||||
label: "codex-dm",
|
||||
).toMatchObject({
|
||||
metadata: expect.objectContaining({
|
||||
pluginBindingOwner: "plugin",
|
||||
pluginId: "openclaw-codex-app-server",
|
||||
pluginRoot: "/Users/huntharo/github/openclaw-app-server",
|
||||
agentId: "codex",
|
||||
boundBy: "system",
|
||||
label: "codex-dm",
|
||||
}),
|
||||
});
|
||||
expect(hoisted.restGet).not.toHaveBeenCalled();
|
||||
expect(hoisted.restPost).not.toHaveBeenCalled();
|
||||
@@ -1430,13 +1346,13 @@ describe("thread binding lifecycle", () => {
|
||||
expect(result.checked).toBe(2);
|
||||
expect(result.removed).toBe(1);
|
||||
expect(result.staleSessionKeys).toContain("agent:codex:acp:stale");
|
||||
expectFields(requireBinding(manager, "thread-acp-healthy"), "healthy binding", {
|
||||
expect(requireBinding(manager, "thread-acp-healthy")).toMatchObject({
|
||||
threadId: "thread-acp-healthy",
|
||||
targetKind: "acp",
|
||||
targetSessionKey: "agent:codex:acp:healthy",
|
||||
});
|
||||
expect(manager.getByThreadId("thread-acp-stale")).toBeUndefined();
|
||||
expectFields(requireBinding(manager, "thread-subagent"), "subagent binding", {
|
||||
expect(requireBinding(manager, "thread-subagent")).toMatchObject({
|
||||
threadId: "thread-subagent",
|
||||
targetKind: "subagent",
|
||||
targetSessionKey: "agent:main:subagent:child",
|
||||
@@ -1482,7 +1398,7 @@ describe("thread binding lifecycle", () => {
|
||||
expect(result.checked).toBe(1);
|
||||
expect(result.removed).toBe(0);
|
||||
expect(result.staleSessionKeys).toStrictEqual([]);
|
||||
expectFields(requireBinding(manager, "thread-acp-uncertain"), "uncertain binding", {
|
||||
expect(requireBinding(manager, "thread-acp-uncertain")).toMatchObject({
|
||||
threadId: "thread-acp-uncertain",
|
||||
targetKind: "acp",
|
||||
targetSessionKey: "agent:codex:acp:uncertain",
|
||||
@@ -1521,16 +1437,12 @@ describe("thread binding lifecycle", () => {
|
||||
expect(result.checked).toBe(0);
|
||||
expect(result.removed).toBe(0);
|
||||
expect(result.staleSessionKeys).toStrictEqual([]);
|
||||
const binding = expectFields(
|
||||
manager.getByThreadId("user:1177378744822943744"),
|
||||
"plugin direct binding",
|
||||
{
|
||||
threadId: "user:1177378744822943744",
|
||||
expect(manager.getByThreadId("user:1177378744822943744")).toMatchObject({
|
||||
threadId: "user:1177378744822943744",
|
||||
metadata: {
|
||||
pluginBindingOwner: "plugin",
|
||||
pluginId: "openclaw-codex-app-server",
|
||||
},
|
||||
);
|
||||
expectFields(requireRecord(binding.metadata, "binding metadata"), "binding metadata", {
|
||||
pluginBindingOwner: "plugin",
|
||||
pluginId: "openclaw-codex-app-server",
|
||||
});
|
||||
});
|
||||
|
||||
@@ -1619,15 +1531,11 @@ describe("thread binding lifecycle", () => {
|
||||
expect(result.checked).toBe(1);
|
||||
expect(result.removed).toBe(0);
|
||||
expect(result.staleSessionKeys).toStrictEqual([]);
|
||||
expectFields(
|
||||
requireBinding(manager, "thread-acp-running-uncertain"),
|
||||
"running uncertain binding",
|
||||
{
|
||||
threadId: "thread-acp-running-uncertain",
|
||||
targetKind: "acp",
|
||||
targetSessionKey: "agent:codex:acp:running-uncertain",
|
||||
},
|
||||
);
|
||||
expect(requireBinding(manager, "thread-acp-running-uncertain")).toMatchObject({
|
||||
threadId: "thread-acp-running-uncertain",
|
||||
targetKind: "acp",
|
||||
targetSessionKey: "agent:codex:acp:running-uncertain",
|
||||
});
|
||||
});
|
||||
|
||||
it("keeps ACP bindings in stored error state when no explicit stale probe verdict exists", async () => {
|
||||
@@ -1670,7 +1578,7 @@ describe("thread binding lifecycle", () => {
|
||||
expect(result.checked).toBe(1);
|
||||
expect(result.removed).toBe(0);
|
||||
expect(result.staleSessionKeys).toStrictEqual([]);
|
||||
expectFields(requireBinding(manager, "thread-acp-error"), "error binding", {
|
||||
expect(requireBinding(manager, "thread-acp-error")).toMatchObject({
|
||||
threadId: "thread-acp-error",
|
||||
targetKind: "acp",
|
||||
targetSessionKey: "agent:codex:acp:error",
|
||||
|
||||
@@ -1,36 +1,19 @@
|
||||
import { Readable } from "node:stream";
|
||||
import { describe, expect, it } from "vitest";
|
||||
import { decodeOpusStream, resolveOpusDecoderPreference } from "./audio.js";
|
||||
import { decodeOpusStream } from "./audio.js";
|
||||
|
||||
describe("discord voice opus decoder selection", () => {
|
||||
it("defaults to the pure-JS opusscript decoder", async () => {
|
||||
it("prefers the pure-JS opusscript decoder over optional native opus", async () => {
|
||||
const verbose: string[] = [];
|
||||
const warnings: string[] = [];
|
||||
const previousPreference = process.env.OPENCLAW_DISCORD_OPUS_DECODER;
|
||||
delete process.env.OPENCLAW_DISCORD_OPUS_DECODER;
|
||||
|
||||
try {
|
||||
const decoded = await decodeOpusStream(Readable.from([]), {
|
||||
onVerbose: (message) => verbose.push(message),
|
||||
onWarn: (message) => warnings.push(message),
|
||||
});
|
||||
const decoded = await decodeOpusStream(Readable.from([]), {
|
||||
onVerbose: (message) => verbose.push(message),
|
||||
onWarn: (message) => warnings.push(message),
|
||||
});
|
||||
|
||||
expect(decoded.length).toBe(0);
|
||||
expect(verbose).toContain("opus decoder: opusscript");
|
||||
expect(warnings).toEqual([]);
|
||||
} finally {
|
||||
if (previousPreference === undefined) {
|
||||
delete process.env.OPENCLAW_DISCORD_OPUS_DECODER;
|
||||
} else {
|
||||
process.env.OPENCLAW_DISCORD_OPUS_DECODER = previousPreference;
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
it("requires an explicit preference for native opus", () => {
|
||||
expect(resolveOpusDecoderPreference()).toBe("opusscript");
|
||||
expect(resolveOpusDecoderPreference("opusscript")).toBe("opusscript");
|
||||
expect(resolveOpusDecoderPreference("native")).toBe("native");
|
||||
expect(resolveOpusDecoderPreference("@discordjs/opus")).toBe("native");
|
||||
expect(decoded.length).toBe(0);
|
||||
expect(verbose).toContain("opus decoder: opusscript");
|
||||
expect(warnings).toEqual([]);
|
||||
});
|
||||
});
|
||||
|
||||
@@ -21,8 +21,6 @@ type OpusDecoderFactory = {
|
||||
name: string;
|
||||
};
|
||||
|
||||
type OpusDecoderPreference = "native" | "opusscript";
|
||||
|
||||
let warnedOpusMissing = false;
|
||||
let cachedOpusDecoderFactory: OpusDecoderFactory | null | "unresolved" = "unresolved";
|
||||
|
||||
@@ -49,34 +47,32 @@ function buildWavBuffer(pcm: Buffer): Buffer {
|
||||
function resolveOpusDecoderFactory(params: {
|
||||
onWarn: (message: string) => void;
|
||||
}): OpusDecoderFactory | null {
|
||||
const nativeFactory: OpusDecoderFactory = {
|
||||
name: "@discordjs/opus",
|
||||
load: () => {
|
||||
const DiscordOpus = require("@discordjs/opus") as {
|
||||
OpusEncoder: new (
|
||||
sampleRate: number,
|
||||
channels: number,
|
||||
) => {
|
||||
decode: (buffer: Buffer) => Buffer;
|
||||
const factories: OpusDecoderFactory[] = [
|
||||
{
|
||||
name: "opusscript",
|
||||
load: () => {
|
||||
const OpusScript = require("opusscript") as {
|
||||
new (sampleRate: number, channels: number, application: number): OpusDecoder;
|
||||
Application: { AUDIO: number };
|
||||
};
|
||||
};
|
||||
return new DiscordOpus.OpusEncoder(SAMPLE_RATE, CHANNELS);
|
||||
return new OpusScript(SAMPLE_RATE, CHANNELS, OpusScript.Application.AUDIO);
|
||||
},
|
||||
},
|
||||
};
|
||||
const opusscriptFactory: OpusDecoderFactory = {
|
||||
name: "opusscript",
|
||||
load: () => {
|
||||
const OpusScript = require("opusscript") as {
|
||||
new (sampleRate: number, channels: number, application: number): OpusDecoder;
|
||||
Application: { AUDIO: number };
|
||||
};
|
||||
return new OpusScript(SAMPLE_RATE, CHANNELS, OpusScript.Application.AUDIO);
|
||||
{
|
||||
name: "@discordjs/opus",
|
||||
load: () => {
|
||||
const DiscordOpus = require("@discordjs/opus") as {
|
||||
OpusEncoder: new (
|
||||
sampleRate: number,
|
||||
channels: number,
|
||||
) => {
|
||||
decode: (buffer: Buffer) => Buffer;
|
||||
};
|
||||
};
|
||||
return new DiscordOpus.OpusEncoder(SAMPLE_RATE, CHANNELS);
|
||||
},
|
||||
},
|
||||
};
|
||||
const factories: OpusDecoderFactory[] =
|
||||
resolveOpusDecoderPreference() === "native"
|
||||
? [nativeFactory, opusscriptFactory]
|
||||
: [opusscriptFactory, nativeFactory];
|
||||
];
|
||||
|
||||
const failures: string[] = [];
|
||||
for (const factory of factories) {
|
||||
@@ -97,16 +93,6 @@ function resolveOpusDecoderFactory(params: {
|
||||
return null;
|
||||
}
|
||||
|
||||
export function resolveOpusDecoderPreference(
|
||||
value = process.env.OPENCLAW_DISCORD_OPUS_DECODER,
|
||||
): OpusDecoderPreference {
|
||||
const normalized = value?.trim().toLowerCase();
|
||||
if (normalized === "native" || normalized === "@discordjs/opus") {
|
||||
return "native";
|
||||
}
|
||||
return "opusscript";
|
||||
}
|
||||
|
||||
function getOrCreateOpusDecoderFactory(params: {
|
||||
onWarn: (message: string) => void;
|
||||
}): OpusDecoderFactory | null {
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
import { PassThrough, type Readable } from "node:stream";
|
||||
import type { Readable } from "node:stream";
|
||||
import { beforeAll, beforeEach, describe, expect, it, vi } from "vitest";
|
||||
import { ChannelType } from "../internal/discord.js";
|
||||
import { createVoiceCaptureState } from "./capture-state.js";
|
||||
@@ -1223,70 +1223,9 @@ describe("DiscordVoiceManager", () => {
|
||||
);
|
||||
});
|
||||
|
||||
it("skips incomplete and non-actionable forced agent-proxy transcripts", async () => {
|
||||
agentCommandMock.mockResolvedValueOnce({ payloads: [{ text: "valid answer" }] });
|
||||
const manager = createManager({
|
||||
groupPolicy: "open",
|
||||
voice: {
|
||||
enabled: true,
|
||||
mode: "agent-proxy",
|
||||
realtime: { provider: "openai" },
|
||||
},
|
||||
});
|
||||
|
||||
await manager.join({ guildId: "g1", channelId: "1001" });
|
||||
const entry = getSessionEntry(manager) as {
|
||||
realtime?: {
|
||||
beginSpeakerTurn: (
|
||||
context: { extraSystemPrompt?: string; senderIsOwner: boolean; speakerLabel: string },
|
||||
userId: string,
|
||||
) => { close: () => void; sendInputAudio: (audio: Buffer) => void };
|
||||
};
|
||||
};
|
||||
const bridgeParams = createRealtimeVoiceBridgeSessionMock.mock.calls.at(-1)?.[0] as
|
||||
| {
|
||||
onTranscript?: (role: "user" | "assistant", text: string, isFinal: boolean) => void;
|
||||
}
|
||||
| undefined;
|
||||
|
||||
const incompleteTurn = entry.realtime?.beginSpeakerTurn(
|
||||
{ extraSystemPrompt: undefined, senderIsOwner: true, speakerLabel: "Owner" },
|
||||
"u-owner",
|
||||
);
|
||||
incompleteTurn?.sendInputAudio(Buffer.alloc(8));
|
||||
bridgeParams?.onTranscript?.("user", "Get this working and...", true);
|
||||
|
||||
const closingTurn = entry.realtime?.beginSpeakerTurn(
|
||||
{ extraSystemPrompt: undefined, senderIsOwner: true, speakerLabel: "Owner" },
|
||||
"u-owner",
|
||||
);
|
||||
closingTurn?.sendInputAudio(Buffer.alloc(8));
|
||||
bridgeParams?.onTranscript?.("user", "I'll be right back. See you guys. Bye-bye.", true);
|
||||
|
||||
await new Promise((resolve) => setTimeout(resolve, 260));
|
||||
expect(agentCommandMock).not.toHaveBeenCalled();
|
||||
|
||||
const validTurn = entry.realtime?.beginSpeakerTurn(
|
||||
{ extraSystemPrompt: undefined, senderIsOwner: true, speakerLabel: "Owner" },
|
||||
"u-owner",
|
||||
);
|
||||
validTurn?.sendInputAudio(Buffer.alloc(8));
|
||||
bridgeParams?.onTranscript?.("user", "ship it.", true);
|
||||
|
||||
await new Promise((resolve) => setTimeout(resolve, 260));
|
||||
expect(agentCommandMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({ message: expect.stringContaining("ship it.") }),
|
||||
expect.anything(),
|
||||
);
|
||||
expect(realtimeSessionMock.sendUserMessage).toHaveBeenCalledWith(
|
||||
expect.stringContaining("valid answer"),
|
||||
);
|
||||
});
|
||||
|
||||
it("queues forced agent-proxy answers until current realtime playback idles", async () => {
|
||||
let resolveFirst: ((value: { payloads: Array<{ text: string }> }) => void) | undefined;
|
||||
let resolveSecond: ((value: { payloads: Array<{ text: string }> }) => void) | undefined;
|
||||
let resolveThird: ((value: { payloads: Array<{ text: string }> }) => void) | undefined;
|
||||
agentCommandMock
|
||||
.mockImplementationOnce(
|
||||
() =>
|
||||
@@ -1299,12 +1238,6 @@ describe("DiscordVoiceManager", () => {
|
||||
new Promise<{ payloads: Array<{ text: string }> }>((resolve) => {
|
||||
resolveSecond = resolve;
|
||||
}),
|
||||
)
|
||||
.mockImplementationOnce(
|
||||
() =>
|
||||
new Promise<{ payloads: Array<{ text: string }> }>((resolve) => {
|
||||
resolveThird = resolve;
|
||||
}),
|
||||
);
|
||||
const manager = createManager({
|
||||
groupPolicy: "open",
|
||||
@@ -1330,7 +1263,6 @@ describe("DiscordVoiceManager", () => {
|
||||
const bridgeParams = createRealtimeVoiceBridgeSessionMock.mock.calls.at(-1)?.[0] as
|
||||
| {
|
||||
audioSink?: { sendAudio: (audio: Buffer) => void };
|
||||
onEvent?: (event: { direction: "server"; type: string }) => void;
|
||||
onTranscript?: (role: "user" | "assistant", text: string, isFinal: boolean) => void;
|
||||
}
|
||||
| undefined;
|
||||
@@ -1347,12 +1279,6 @@ describe("DiscordVoiceManager", () => {
|
||||
);
|
||||
secondTurn?.sendInputAudio(Buffer.alloc(8));
|
||||
bridgeParams?.onTranscript?.("user", "second question", true);
|
||||
const thirdTurn = entry.realtime?.beginSpeakerTurn(
|
||||
{ extraSystemPrompt: undefined, senderIsOwner: true, speakerLabel: "Owner" },
|
||||
"u-owner",
|
||||
);
|
||||
thirdTurn?.sendInputAudio(Buffer.alloc(8));
|
||||
bridgeParams?.onTranscript?.("user", "third question", true);
|
||||
await new Promise((resolve) => setTimeout(resolve, 260));
|
||||
|
||||
resolveFirst?.({ payloads: [{ text: "first answer" }] });
|
||||
@@ -1364,18 +1290,6 @@ describe("DiscordVoiceManager", () => {
|
||||
bridgeParams?.audioSink?.sendAudio(Buffer.alloc(480));
|
||||
|
||||
resolveSecond?.({ payloads: [{ text: "second answer" }] });
|
||||
resolveThird?.({ payloads: [{ text: "third answer" }] });
|
||||
await new Promise((resolve) => setTimeout(resolve, 0));
|
||||
expect(realtimeSessionMock.sendUserMessage).not.toHaveBeenCalledWith(
|
||||
expect.stringContaining("second answer"),
|
||||
);
|
||||
expect(realtimeSessionMock.sendUserMessage).not.toHaveBeenCalledWith(
|
||||
expect.stringContaining("third answer"),
|
||||
);
|
||||
|
||||
bridgeParams?.onEvent?.({ direction: "server", type: "response.done" });
|
||||
const firstStream = createAudioResourceMock.mock.calls.at(-1)?.[0] as PassThrough | undefined;
|
||||
await vi.waitFor(() => expect(firstStream?.writableEnded).toBe(true));
|
||||
await new Promise((resolve) => setTimeout(resolve, 0));
|
||||
expect(realtimeSessionMock.sendUserMessage).not.toHaveBeenCalledWith(
|
||||
expect.stringContaining("second answer"),
|
||||
@@ -1388,23 +1302,6 @@ describe("DiscordVoiceManager", () => {
|
||||
expect(realtimeSessionMock.sendUserMessage).toHaveBeenCalledWith(
|
||||
expect.stringContaining("second answer"),
|
||||
);
|
||||
expect(realtimeSessionMock.sendUserMessage).not.toHaveBeenCalledWith(
|
||||
expect.stringContaining("third answer"),
|
||||
);
|
||||
|
||||
bridgeParams?.audioSink?.sendAudio(Buffer.alloc(480));
|
||||
bridgeParams?.onEvent?.({ direction: "server", type: "response.done" });
|
||||
const secondStream = createAudioResourceMock.mock.calls.at(-1)?.[0] as PassThrough | undefined;
|
||||
await vi.waitFor(() => expect(secondStream?.writableEnded).toBe(true));
|
||||
await new Promise((resolve) => setTimeout(resolve, 0));
|
||||
expect(realtimeSessionMock.sendUserMessage).not.toHaveBeenCalledWith(
|
||||
expect.stringContaining("third answer"),
|
||||
);
|
||||
|
||||
idleHandler?.();
|
||||
expect(realtimeSessionMock.sendUserMessage).toHaveBeenCalledWith(
|
||||
expect.stringContaining("third answer"),
|
||||
);
|
||||
});
|
||||
|
||||
it("matches agent-proxy consult tool calls to the pending transcript", async () => {
|
||||
|
||||
@@ -46,35 +46,6 @@ const DISCORD_REALTIME_LOG_PREVIEW_CHARS = 500;
|
||||
const DISCORD_REALTIME_DEFAULT_MIN_BARGE_IN_AUDIO_END_MS = 250;
|
||||
const DISCORD_REALTIME_FORCED_CONSULT_FALLBACK_DELAY_MS = 200;
|
||||
const REALTIME_PCM16_BYTES_PER_SAMPLE = 2;
|
||||
const DISCORD_REALTIME_FORCED_CONSULT_TRAILING_FRAGMENT_WORDS = new Set([
|
||||
"a",
|
||||
"about",
|
||||
"an",
|
||||
"and",
|
||||
"as",
|
||||
"at",
|
||||
"because",
|
||||
"but",
|
||||
"by",
|
||||
"for",
|
||||
"from",
|
||||
"in",
|
||||
"of",
|
||||
"on",
|
||||
"or",
|
||||
"so",
|
||||
"that",
|
||||
"the",
|
||||
"then",
|
||||
"to",
|
||||
"with",
|
||||
]);
|
||||
const DISCORD_REALTIME_VERBOSE_OMITTED_EVENTS = new Set([
|
||||
"conversation.output_audio.delta",
|
||||
"input_audio_buffer.append",
|
||||
"response.audio.delta",
|
||||
"response.output_audio.delta",
|
||||
]);
|
||||
|
||||
export type DiscordVoiceMode = "stt-tts" | "agent-proxy" | "bidi";
|
||||
|
||||
@@ -152,32 +123,6 @@ function formatRealtimeInterruptionLog(event: RealtimeVoiceBridgeEvent): string
|
||||
return undefined;
|
||||
}
|
||||
|
||||
function shouldLogRealtimeVerboseEvent(event: RealtimeVoiceBridgeEvent): boolean {
|
||||
return !DISCORD_REALTIME_VERBOSE_OMITTED_EVENTS.has(event.type);
|
||||
}
|
||||
|
||||
function classifySkippableForcedAgentProxyTranscript(text: string): string | undefined {
|
||||
const normalized = text.replace(/\s+/g, " ").trim().toLowerCase();
|
||||
if (!normalized) {
|
||||
return "empty";
|
||||
}
|
||||
if (/(\.\.\.|…)\s*$/.test(normalized)) {
|
||||
return "incomplete-transcript";
|
||||
}
|
||||
const lastWord = normalized.match(/[a-z']+$/)?.[0]?.replace(/^'+|'+$/g, "");
|
||||
if (lastWord && DISCORD_REALTIME_FORCED_CONSULT_TRAILING_FRAGMENT_WORDS.has(lastWord)) {
|
||||
return "trailing-fragment";
|
||||
}
|
||||
if (
|
||||
!normalized.includes("?") &&
|
||||
(/^(i'?ll|i will) be (right )?back\b/.test(normalized) ||
|
||||
/\b(see you|bye(?:-bye)?|goodbye)\b/.test(normalized))
|
||||
) {
|
||||
return "non-actionable-closing";
|
||||
}
|
||||
return undefined;
|
||||
}
|
||||
|
||||
function readProviderConfigString(
|
||||
config: RealtimeVoiceProviderConfig,
|
||||
key: string,
|
||||
@@ -436,9 +381,7 @@ export class DiscordRealtimeVoiceSession implements VoiceRealtimeSession {
|
||||
onToolCall: (event, session) => this.handleToolCall(event, session),
|
||||
onEvent: (event) => {
|
||||
const detail = event.detail ? ` ${event.detail}` : "";
|
||||
if (shouldLogRealtimeVerboseEvent(event)) {
|
||||
logVoiceVerbose(`realtime ${event.direction}:${event.type}${detail}`);
|
||||
}
|
||||
logVoiceVerbose(`realtime ${event.direction}:${event.type}${detail}`);
|
||||
const responseEnded =
|
||||
event.direction === "server" &&
|
||||
(event.type === "response.done" || event.type === "response.cancelled");
|
||||
@@ -596,12 +539,6 @@ export class DiscordRealtimeVoiceSession implements VoiceRealtimeSession {
|
||||
return;
|
||||
}
|
||||
this.syncOutputAudioTimestamp();
|
||||
if (this.outputStreamEnding) {
|
||||
logVoiceVerbose(
|
||||
`realtime output audio ignored after stream ending: guild ${this.params.entry.guildId} channel ${this.params.entry.channelId}`,
|
||||
);
|
||||
return;
|
||||
}
|
||||
const stream = this.ensureOutputStream();
|
||||
if (this.exactSpeechResponseActive) {
|
||||
this.exactSpeechAudioStarted = true;
|
||||
@@ -617,7 +554,7 @@ export class DiscordRealtimeVoiceSession implements VoiceRealtimeSession {
|
||||
}
|
||||
|
||||
private ensureOutputStream(): PassThrough {
|
||||
if (this.outputStream && !this.outputStream.destroyed && !this.outputStream.writableEnded) {
|
||||
if (this.outputStream && !this.outputStream.destroyed) {
|
||||
return this.outputStream;
|
||||
}
|
||||
const voiceSdk = loadDiscordVoiceSdk();
|
||||
@@ -629,7 +566,7 @@ export class DiscordRealtimeVoiceSession implements VoiceRealtimeSession {
|
||||
this.logOutputAudioStopped("stream-close");
|
||||
this.outputStream = null;
|
||||
this.resetOutputAudioStats();
|
||||
this.completeExactSpeechResponse("stream-close", { drain: false });
|
||||
this.completeExactSpeechResponse("stream-close");
|
||||
}
|
||||
});
|
||||
const resource = voiceSdk.createAudioResource(stream, {
|
||||
@@ -692,15 +629,12 @@ export class DiscordRealtimeVoiceSession implements VoiceRealtimeSession {
|
||||
this.bridge?.sendUserMessage(buildDiscordSpeakExactUserMessage(text));
|
||||
}
|
||||
|
||||
private completeExactSpeechResponse(reason: string, options?: { drain?: boolean }): void {
|
||||
private completeExactSpeechResponse(reason: string): void {
|
||||
if (!this.exactSpeechResponseActive && this.queuedExactSpeechMessages.length === 0) {
|
||||
return;
|
||||
}
|
||||
this.exactSpeechResponseActive = false;
|
||||
this.exactSpeechAudioStarted = false;
|
||||
if (options?.drain === false) {
|
||||
return;
|
||||
}
|
||||
this.drainQueuedExactSpeechMessages(reason);
|
||||
}
|
||||
|
||||
@@ -877,13 +811,6 @@ export class DiscordRealtimeVoiceSession implements VoiceRealtimeSession {
|
||||
return;
|
||||
}
|
||||
const context = this.consumePendingSpeakerContext();
|
||||
const skipReason = classifySkippableForcedAgentProxyTranscript(question);
|
||||
if (skipReason) {
|
||||
logger.info(
|
||||
`discord voice: realtime forced agent consult skipped reason=${skipReason} chars=${question.length} speaker=${context?.speakerLabel ?? "unknown"} transcript=${formatRealtimeLogPreview(question)}`,
|
||||
);
|
||||
return;
|
||||
}
|
||||
if (!context) {
|
||||
const recent = this.findRecentAgentProxyConsultContext(question);
|
||||
if (recent) {
|
||||
|
||||
@@ -123,51 +123,6 @@ describe("Feishu Card Action Handler", () => {
|
||||
resetProcessedFeishuCardActionTokensForTests();
|
||||
});
|
||||
|
||||
function mockCallArg(
|
||||
mock: { mock: { calls: unknown[][] } },
|
||||
index: number,
|
||||
label: string,
|
||||
): unknown {
|
||||
const call = mock.mock.calls[index];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} call ${index + 1}`);
|
||||
}
|
||||
return call[0];
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object") {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function handleMessageEvent(callIndex = 0) {
|
||||
const arg = requireRecord(
|
||||
mockCallArg(vi.mocked(handleFeishuMessage), callIndex, "handleFeishuMessage"),
|
||||
"handleFeishuMessage args",
|
||||
);
|
||||
return requireRecord(arg.event, "Feishu message event");
|
||||
}
|
||||
|
||||
function handleMessage(callIndex = 0) {
|
||||
return requireRecord(handleMessageEvent(callIndex).message, "Feishu message");
|
||||
}
|
||||
|
||||
function sendMessageCall(callIndex = 0) {
|
||||
return requireRecord(
|
||||
mockCallArg(sendMessageFeishuMock, callIndex, "sendMessageFeishu"),
|
||||
"sendMessageFeishu args",
|
||||
);
|
||||
}
|
||||
|
||||
function sendCardCall(callIndex = 0) {
|
||||
return requireRecord(
|
||||
mockCallArg(sendCardFeishuMock, callIndex, "sendCardFeishu"),
|
||||
"sendCardFeishu args",
|
||||
);
|
||||
}
|
||||
|
||||
it("handles card action with text payload", async () => {
|
||||
const event: FeishuCardActionEvent = {
|
||||
operator: { open_id: "u123", user_id: "uid1", union_id: "un1" },
|
||||
@@ -178,9 +133,16 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const message = handleMessage();
|
||||
expect(message.content).toBe('{"text":"/ping"}');
|
||||
expect(message.chat_id).toBe("chat1");
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
content: '{"text":"/ping"}',
|
||||
chat_id: "chat1",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("handles card action with JSON object payload", async () => {
|
||||
@@ -193,9 +155,16 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const message = handleMessage();
|
||||
expect(message.content).toBe('{"text":"{\\"key\\":\\"val\\"}"}');
|
||||
expect(message.chat_id).toBe("u123"); // Fallback to open_id
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
content: '{"text":"{\\"key\\":\\"val\\"}"}',
|
||||
chat_id: "u123", // Fallback to open_id
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("routes quick command actions with operator and conversation context", async () => {
|
||||
@@ -207,15 +176,23 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const eventArg = handleMessageEvent();
|
||||
const sender = requireRecord(eventArg.sender, "Feishu sender");
|
||||
const senderId = requireRecord(sender.sender_id, "Feishu sender id");
|
||||
expect(senderId.open_id).toBe("u123");
|
||||
expect(senderId.user_id).toBe("uid1");
|
||||
expect(senderId.union_id).toBe("un1");
|
||||
const message = requireRecord(eventArg.message, "Feishu message");
|
||||
expect(message.chat_id).toBe("chat1");
|
||||
expect(message.content).toBe('{"text":"/help"}');
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
sender: expect.objectContaining({
|
||||
sender_id: expect.objectContaining({
|
||||
open_id: "u123",
|
||||
user_id: "uid1",
|
||||
union_id: "un1",
|
||||
}),
|
||||
}),
|
||||
message: expect.objectContaining({
|
||||
chat_id: "chat1",
|
||||
content: '{"text":"/help"}',
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("opens an approval card for metadata actions", async () => {
|
||||
@@ -245,27 +222,39 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime, accountId: "main" });
|
||||
|
||||
const cardCall = sendCardCall();
|
||||
expect(cardCall.to).toBe("chat:chat1");
|
||||
expect(cardCall.accountId).toBe("main");
|
||||
const card = requireRecord(cardCall.card, "Feishu card");
|
||||
expect(requireRecord(card.config, "Feishu card config").width_mode).toBe("fill");
|
||||
const header = requireRecord(card.header, "Feishu card header");
|
||||
expect(requireRecord(header.title, "Feishu card title").content).toBe("Confirm action");
|
||||
const body = requireRecord(card.body, "Feishu card body");
|
||||
const elements = body.elements as Array<Record<string, unknown>>;
|
||||
const actionElement = elements.find((element) => element.tag === "action");
|
||||
if (!actionElement) {
|
||||
throw new Error("Expected action element");
|
||||
}
|
||||
const actions = actionElement.actions as Array<Record<string, unknown>>;
|
||||
const actionValue = requireRecord(actions[0]?.value, "Feishu approval action value");
|
||||
const approvalContext = requireRecord(actionValue.c, "Feishu approval context");
|
||||
expect(approvalContext.u).toBe("u123");
|
||||
expect(approvalContext.h).toBe("chat1");
|
||||
expect(approvalContext.t).toBe("group");
|
||||
expect(approvalContext.s).toBe("agent:codex:feishu:chat:chat1");
|
||||
expect(typeof approvalContext.e).toBe("number");
|
||||
expect(sendCardFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
to: "chat:chat1",
|
||||
accountId: "main",
|
||||
card: expect.objectContaining({
|
||||
config: expect.objectContaining({
|
||||
width_mode: "fill",
|
||||
}),
|
||||
header: expect.objectContaining({
|
||||
title: expect.objectContaining({ content: "Confirm action" }),
|
||||
}),
|
||||
body: expect.objectContaining({
|
||||
elements: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
tag: "action",
|
||||
actions: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
value: expect.objectContaining({
|
||||
c: expect.objectContaining({
|
||||
u: "u123",
|
||||
h: "chat1",
|
||||
t: "group",
|
||||
s: "agent:codex:feishu:chat:chat1",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectFirstSentCardUsesFillWidthOnly(sendCardFeishuMock);
|
||||
expect(handleFeishuMessage).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -279,7 +268,15 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
expect(handleMessage().content).toBe('{"text":"/new"}');
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
content: '{"text":"/new"}',
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("safely rejects stale structured actions", async () => {
|
||||
@@ -295,9 +292,12 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const sendMessage = sendMessageCall();
|
||||
expect(sendMessage.to).toBe("chat:chat1");
|
||||
expect(String(sendMessage.text)).toContain("expired");
|
||||
expect(sendMessageFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
to: "chat:chat1",
|
||||
text: expect.stringContaining("expired"),
|
||||
}),
|
||||
);
|
||||
expect(handleFeishuMessage).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
@@ -312,7 +312,11 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
expect(String(sendMessageCall().text)).toContain("different user");
|
||||
expect(sendMessageFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
text: expect.stringContaining("different user"),
|
||||
}),
|
||||
);
|
||||
expect(handleFeishuMessage).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
@@ -333,9 +337,12 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const sendMessage = sendMessageCall();
|
||||
expect(sendMessage.to).toBe("chat:chat1");
|
||||
expect(sendMessage.text).toBe("Cancelled.");
|
||||
expect(sendMessageFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
to: "chat:chat1",
|
||||
text: "Cancelled.",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("preserves p2p callbacks for DM quick actions", async () => {
|
||||
@@ -349,9 +356,16 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const message = handleMessage();
|
||||
expect(message.chat_id).toBe("p2p-chat-1");
|
||||
expect(message.chat_type).toBe("p2p");
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
chat_id: "p2p-chat-1",
|
||||
chat_type: "p2p",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("resolves DM chat type from the Feishu chat API when card context omits it", async () => {
|
||||
@@ -370,9 +384,16 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
const message = handleMessage();
|
||||
expect(message.chat_id).toBe("oc_dm_chat_123");
|
||||
expect(message.chat_type).toBe("p2p");
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
chat_id: "oc_dm_chat_123",
|
||||
chat_type: "p2p",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(createFeishuClientMock).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
@@ -424,7 +445,15 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
expect(handleMessage().chat_type).toBe("p2p");
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
chat_type: "p2p",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("falls back to p2p when Feishu chat API throws", async () => {
|
||||
@@ -443,7 +472,15 @@ describe("Feishu Card Action Handler", () => {
|
||||
|
||||
await handleFeishuCardAction({ cfg, event, runtime });
|
||||
|
||||
expect(handleMessage().chat_type).toBe("p2p");
|
||||
expect(handleFeishuMessage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
event: expect.objectContaining({
|
||||
message: expect.objectContaining({
|
||||
chat_type: "p2p",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("drops duplicate structured callback tokens", async () => {
|
||||
|
||||
@@ -59,36 +59,6 @@ function getDescribedActions(cfg: OpenClawConfig, accountId?: string): string[]
|
||||
return [...(feishuPlugin.actions?.describeMessageTool?.({ cfg, accountId })?.actions ?? [])];
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): unknown[] {
|
||||
if (!Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value;
|
||||
}
|
||||
|
||||
function mockCallArg(mock: unknown, callIndex: number, argIndex: number, label: string) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`Expected ${label} mock calls`);
|
||||
}
|
||||
const call = calls[callIndex];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} call ${callIndex + 1}`);
|
||||
}
|
||||
return call[argIndex];
|
||||
}
|
||||
|
||||
function resultDetails(result: unknown) {
|
||||
return requireRecord(requireRecord(result, "action result").details, "action result details");
|
||||
}
|
||||
|
||||
afterAll(() => {
|
||||
vi.doUnmock("./probe.js");
|
||||
vi.doUnmock("./client.js");
|
||||
@@ -123,16 +93,14 @@ describe("feishuPlugin.status.probeAccount", () => {
|
||||
});
|
||||
|
||||
expect(probeFeishuMock).toHaveBeenCalledTimes(1);
|
||||
const probeArgs = requireRecord(
|
||||
mockCallArg(probeFeishuMock, 0, 0, "probeFeishu"),
|
||||
"probe args",
|
||||
expect(probeFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
accountId: "main",
|
||||
appId: "cli_main",
|
||||
appSecret: "secret_main",
|
||||
}),
|
||||
);
|
||||
expect(probeArgs.accountId).toBe("main");
|
||||
expect(probeArgs.appId).toBe("cli_main");
|
||||
expect(probeArgs.appSecret).toBe("secret_main");
|
||||
const resultRecord = requireRecord(result, "probe result");
|
||||
expect(resultRecord.ok).toBe(true);
|
||||
expect(resultRecord.appId).toBe("cli_main");
|
||||
expect(result).toMatchObject({ ok: true, appId: "cli_main" });
|
||||
});
|
||||
});
|
||||
|
||||
@@ -163,13 +131,13 @@ describe("feishuPlugin.pairing.notifyApproval", () => {
|
||||
accountId: "work",
|
||||
});
|
||||
|
||||
const sendArgs = requireRecord(
|
||||
mockCallArg(sendMessageFeishuMock, 0, 0, "sendMessageFeishu"),
|
||||
"send args",
|
||||
expect(sendMessageFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
cfg,
|
||||
to: "ou_user",
|
||||
accountId: "work",
|
||||
}),
|
||||
);
|
||||
expect(sendArgs.cfg).toBe(cfg);
|
||||
expect(sendArgs.to).toBe("ou_user");
|
||||
expect(sendArgs.accountId).toBe("work");
|
||||
});
|
||||
});
|
||||
|
||||
@@ -343,10 +311,7 @@ describe("feishuPlugin actions", () => {
|
||||
replyToMessageId: undefined,
|
||||
replyInThread: false,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(details.messageId).toBe("om_sent");
|
||||
expect(details.chatId).toBe("oc_group_1");
|
||||
expect(result?.details).toMatchObject({ ok: true, messageId: "om_sent", chatId: "oc_group_1" });
|
||||
});
|
||||
|
||||
it("renders presentation messages as cards", async () => {
|
||||
@@ -366,33 +331,29 @@ describe("feishuPlugin actions", () => {
|
||||
toolContext: {},
|
||||
} as never);
|
||||
|
||||
const sendCardArgs = requireRecord(
|
||||
mockCallArg(sendCardFeishuMock, 0, 0, "sendCardFeishu"),
|
||||
"send card args",
|
||||
);
|
||||
expect(sendCardArgs.cfg).toBe(cfg);
|
||||
expect(sendCardArgs.to).toBe("chat:oc_group_1");
|
||||
expect(sendCardArgs.accountId).toBeUndefined();
|
||||
expect(sendCardArgs.replyToMessageId).toBeUndefined();
|
||||
expect(sendCardArgs.replyInThread).toBe(false);
|
||||
const card = requireRecord(sendCardArgs.card, "card");
|
||||
expect(card.schema).toBe("2.0");
|
||||
expect(card.header).toEqual({
|
||||
title: { tag: "plain_text", content: "Status" },
|
||||
template: "blue",
|
||||
});
|
||||
expect(card.body).toEqual({
|
||||
elements: [
|
||||
{
|
||||
tag: "markdown",
|
||||
content: "Build completed",
|
||||
expect(sendCardFeishuMock).toHaveBeenCalledWith({
|
||||
cfg,
|
||||
to: "chat:oc_group_1",
|
||||
card: expect.objectContaining({
|
||||
schema: "2.0",
|
||||
header: {
|
||||
title: { tag: "plain_text", content: "Status" },
|
||||
template: "blue",
|
||||
},
|
||||
],
|
||||
body: {
|
||||
elements: [
|
||||
{
|
||||
tag: "markdown",
|
||||
content: "Build completed",
|
||||
},
|
||||
],
|
||||
},
|
||||
}),
|
||||
accountId: undefined,
|
||||
replyToMessageId: undefined,
|
||||
replyInThread: false,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(details.messageId).toBe("om_card");
|
||||
expect(details.chatId).toBe("oc_group_1");
|
||||
expect(result?.details).toMatchObject({ ok: true, messageId: "om_card", chatId: "oc_group_1" });
|
||||
});
|
||||
|
||||
it("renders presentation button labels into the card fallback", async () => {
|
||||
@@ -416,17 +377,20 @@ describe("feishuPlugin actions", () => {
|
||||
toolContext: {},
|
||||
} as never);
|
||||
|
||||
const sendCardArgs = requireRecord(
|
||||
mockCallArg(sendCardFeishuMock, 0, 0, "sendCardFeishu"),
|
||||
"send card args",
|
||||
expect(sendCardFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
card: expect.objectContaining({
|
||||
body: {
|
||||
elements: [
|
||||
{
|
||||
tag: "markdown",
|
||||
content: "- Run help",
|
||||
},
|
||||
],
|
||||
},
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const card = requireRecord(sendCardArgs.card, "card");
|
||||
expect(requireRecord(card.body, "card body").elements).toEqual([
|
||||
{
|
||||
tag: "markdown",
|
||||
content: "- Run help",
|
||||
},
|
||||
]);
|
||||
});
|
||||
|
||||
it("renders presentation select labels into the card fallback", async () => {
|
||||
@@ -451,17 +415,20 @@ describe("feishuPlugin actions", () => {
|
||||
toolContext: {},
|
||||
} as never);
|
||||
|
||||
const sendCardArgs = requireRecord(
|
||||
mockCallArg(sendCardFeishuMock, 0, 0, "sendCardFeishu"),
|
||||
"send card args",
|
||||
expect(sendCardFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
card: expect.objectContaining({
|
||||
body: {
|
||||
elements: [
|
||||
{
|
||||
tag: "markdown",
|
||||
content: "Pick one:\n- Option A",
|
||||
},
|
||||
],
|
||||
},
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const card = requireRecord(sendCardArgs.card, "card");
|
||||
expect(requireRecord(card.body, "card body").elements).toEqual([
|
||||
{
|
||||
tag: "markdown",
|
||||
content: "Pick one:\n- Option A",
|
||||
},
|
||||
]);
|
||||
});
|
||||
|
||||
it("sends media through the outbound adapter", async () => {
|
||||
@@ -493,7 +460,7 @@ describe("feishuPlugin actions", () => {
|
||||
mediaLocalRoots: ["/tmp"],
|
||||
replyToId: undefined,
|
||||
});
|
||||
expect(resultDetails(result).messageId).toBe("om_media");
|
||||
expect(result?.details).toMatchObject({ messageId: "om_media" });
|
||||
});
|
||||
|
||||
it("passes asVoice through media sends", async () => {
|
||||
@@ -516,12 +483,12 @@ describe("feishuPlugin actions", () => {
|
||||
mediaLocalRoots: [],
|
||||
} as never);
|
||||
|
||||
const mediaArgs = requireRecord(
|
||||
mockCallArg(feishuOutboundSendMediaMock, 0, 0, "feishuOutbound.sendMedia"),
|
||||
"media args",
|
||||
expect(feishuOutboundSendMediaMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
mediaUrl: "https://example.com/reply.mp3",
|
||||
audioAsVoice: true,
|
||||
}),
|
||||
);
|
||||
expect(mediaArgs.mediaUrl).toBe("https://example.com/reply.mp3");
|
||||
expect(mediaArgs.audioAsVoice).toBe(true);
|
||||
});
|
||||
|
||||
it("reads messages", async () => {
|
||||
@@ -543,11 +510,10 @@ describe("feishuPlugin actions", () => {
|
||||
messageId: "om_1",
|
||||
accountId: undefined,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
const message = requireRecord(details.message, "read message");
|
||||
expect(message.messageId).toBe("om_1");
|
||||
expect(message.content).toBe("hello");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
message: expect.objectContaining({ messageId: "om_1", content: "hello" }),
|
||||
});
|
||||
});
|
||||
|
||||
it("returns an error result when message reads fail", async () => {
|
||||
@@ -583,10 +549,7 @@ describe("feishuPlugin actions", () => {
|
||||
card: undefined,
|
||||
accountId: undefined,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(details.messageId).toBe("om_2");
|
||||
expect(details.contentType).toBe("post");
|
||||
expect(result?.details).toMatchObject({ ok: true, messageId: "om_2", contentType: "post" });
|
||||
});
|
||||
|
||||
it("sends explicit thread replies with reply_in_thread semantics", async () => {
|
||||
@@ -608,10 +571,11 @@ describe("feishuPlugin actions", () => {
|
||||
replyToMessageId: "om_parent",
|
||||
replyInThread: true,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(details.action).toBe("thread-reply");
|
||||
expect(details.messageId).toBe("om_reply");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
action: "thread-reply",
|
||||
messageId: "om_reply",
|
||||
});
|
||||
});
|
||||
|
||||
it("auto-threads `send` text against the inbound trigger in group_topic sessions", async () => {
|
||||
@@ -654,12 +618,12 @@ describe("feishuPlugin actions", () => {
|
||||
toolContext: { currentMessageId: "om_inbound" },
|
||||
} as never);
|
||||
|
||||
const sendCardArgs = requireRecord(
|
||||
mockCallArg(sendCardFeishuMock, 0, 0, "sendCardFeishu"),
|
||||
"send card args",
|
||||
expect(sendCardFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
replyToMessageId: "om_inbound",
|
||||
replyInThread: true,
|
||||
}),
|
||||
);
|
||||
expect(sendCardArgs.replyToMessageId).toBe("om_inbound");
|
||||
expect(sendCardArgs.replyInThread).toBe(true);
|
||||
});
|
||||
|
||||
it("auto-threads `send` media against the inbound trigger in group_topic sessions", async () => {
|
||||
@@ -683,12 +647,14 @@ describe("feishuPlugin actions", () => {
|
||||
mediaLocalRoots: ["/tmp"],
|
||||
} as never);
|
||||
|
||||
const mediaArgs = requireRecord(
|
||||
mockCallArg(feishuOutboundSendMediaMock, 0, 0, "feishuOutbound.sendMedia"),
|
||||
"media args",
|
||||
expect(feishuOutboundSendMediaMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
threadId: "om_inbound",
|
||||
}),
|
||||
);
|
||||
expect(feishuOutboundSendMediaMock).toHaveBeenCalledWith(
|
||||
expect.not.objectContaining({ replyToId: expect.anything() }),
|
||||
);
|
||||
expect(mediaArgs.threadId).toBe("om_inbound");
|
||||
expect("replyToId" in mediaArgs).toBe(false);
|
||||
});
|
||||
|
||||
it("auto-threads `send` in group_topic_sender sessions too", async () => {
|
||||
@@ -703,12 +669,12 @@ describe("feishuPlugin actions", () => {
|
||||
toolContext: { currentMessageId: "om_inbound" },
|
||||
} as never);
|
||||
|
||||
const sendArgs = requireRecord(
|
||||
mockCallArg(sendMessageFeishuMock, 0, 0, "sendMessageFeishu"),
|
||||
"send args",
|
||||
expect(sendMessageFeishuMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
replyToMessageId: "om_inbound",
|
||||
replyInThread: true,
|
||||
}),
|
||||
);
|
||||
expect(sendArgs.replyToMessageId).toBe("om_inbound");
|
||||
expect(sendArgs.replyInThread).toBe(true);
|
||||
});
|
||||
|
||||
it("does not auto-thread `send` in plain group sessions (no topic)", async () => {
|
||||
@@ -770,9 +736,10 @@ describe("feishuPlugin actions", () => {
|
||||
messageId: "om_pin",
|
||||
accountId: undefined,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(requireRecord(details.pin, "pin").messageId).toBe("om_pin");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
pin: expect.objectContaining({ messageId: "om_pin" }),
|
||||
});
|
||||
});
|
||||
|
||||
it("lists pins", async () => {
|
||||
@@ -800,10 +767,10 @@ describe("feishuPlugin actions", () => {
|
||||
pageToken: undefined,
|
||||
accountId: undefined,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
const pins = requireArray(details.pins, "pins");
|
||||
expect(requireRecord(pins[0], "pin").messageId).toBe("om_pin");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
pins: [expect.objectContaining({ messageId: "om_pin" })],
|
||||
});
|
||||
});
|
||||
|
||||
it("removes pins", async () => {
|
||||
@@ -819,9 +786,7 @@ describe("feishuPlugin actions", () => {
|
||||
messageId: "om_pin",
|
||||
accountId: undefined,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(details.messageId).toBe("om_pin");
|
||||
expect(result?.details).toMatchObject({ ok: true, messageId: "om_pin" });
|
||||
});
|
||||
|
||||
it("fetches channel info", async () => {
|
||||
@@ -837,11 +802,10 @@ describe("feishuPlugin actions", () => {
|
||||
|
||||
expect(createFeishuClientMock).toHaveBeenCalled();
|
||||
expect(getChatInfoMock).toHaveBeenCalledWith({ tag: "client" }, "oc_group_1");
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
const channel = requireRecord(details.channel, "channel");
|
||||
expect(channel.chat_id).toBe("oc_group_1");
|
||||
expect(channel.name).toBe("Eng");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
channel: expect.objectContaining({ chat_id: "oc_group_1", name: "Eng" }),
|
||||
});
|
||||
});
|
||||
|
||||
it("fetches member lists from a chat", async () => {
|
||||
@@ -866,12 +830,10 @@ describe("feishuPlugin actions", () => {
|
||||
undefined,
|
||||
"open_id",
|
||||
);
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
const members = requireArray(details.members, "members");
|
||||
const member = requireRecord(members[0], "member");
|
||||
expect(member.member_id).toBe("ou_1");
|
||||
expect(member.name).toBe("Alice");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
members: [expect.objectContaining({ member_id: "ou_1", name: "Alice" })],
|
||||
});
|
||||
});
|
||||
|
||||
it("fetches individual member info", async () => {
|
||||
@@ -886,11 +848,10 @@ describe("feishuPlugin actions", () => {
|
||||
} as never);
|
||||
|
||||
expect(getFeishuMemberInfoMock).toHaveBeenCalledWith({ tag: "client" }, "ou_1", "open_id");
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
const member = requireRecord(details.member, "member");
|
||||
expect(member.member_id).toBe("ou_1");
|
||||
expect(member.name).toBe("Alice");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
member: expect.objectContaining({ member_id: "ou_1", name: "Alice" }),
|
||||
});
|
||||
});
|
||||
|
||||
it("infers user_id lookups from the userId alias", async () => {
|
||||
@@ -946,12 +907,11 @@ describe("feishuPlugin actions", () => {
|
||||
fallbackToStatic: false,
|
||||
accountId: undefined,
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
const groups = requireArray(details.groups, "groups");
|
||||
const peers = requireArray(details.peers, "peers");
|
||||
expect(requireRecord(groups[0], "group").id).toBe("oc_group_1");
|
||||
expect(requireRecord(peers[0], "peer").id).toBe("ou_1");
|
||||
expect(result?.details).toMatchObject({
|
||||
ok: true,
|
||||
groups: [expect.objectContaining({ id: "oc_group_1" })],
|
||||
peers: [expect.objectContaining({ id: "ou_1" })],
|
||||
});
|
||||
});
|
||||
|
||||
it("fails channel-list when live discovery fails", async () => {
|
||||
@@ -999,9 +959,7 @@ describe("feishuPlugin actions", () => {
|
||||
accountId: undefined,
|
||||
});
|
||||
expect(removeReactionFeishuMock).toHaveBeenCalledTimes(2);
|
||||
const details = resultDetails(result);
|
||||
expect(details.ok).toBe(true);
|
||||
expect(details.removed).toBe(2);
|
||||
expect(result?.details).toMatchObject({ ok: true, removed: 2 });
|
||||
});
|
||||
|
||||
it("fails for missing params on supported actions", async () => {
|
||||
@@ -1034,13 +992,13 @@ describe("feishuPlugin actions", () => {
|
||||
mediaLocalRoots: [],
|
||||
} as never);
|
||||
|
||||
const mediaArgs = requireRecord(
|
||||
mockCallArg(feishuOutboundSendMediaMock, 0, 0, "feishuOutbound.sendMedia"),
|
||||
"media args",
|
||||
expect(feishuOutboundSendMediaMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
to: "chat:oc_group_1",
|
||||
mediaUrl: "https://example.com/image.png",
|
||||
}),
|
||||
);
|
||||
expect(mediaArgs.to).toBe("chat:oc_group_1");
|
||||
expect(mediaArgs.mediaUrl).toBe("https://example.com/image.png");
|
||||
expect(resultDetails(result).messageId).toBe("om_media_only");
|
||||
expect(result?.details).toMatchObject({ messageId: "om_media_only" });
|
||||
});
|
||||
|
||||
it("fails for unsupported action names", async () => {
|
||||
|
||||
@@ -115,57 +115,6 @@ function buildDelayedSecondSseResponse(params: {
|
||||
});
|
||||
}
|
||||
|
||||
function requireMockCall<TArgs extends unknown[]>(
|
||||
mock: { mock: { calls: TArgs[] } },
|
||||
index: number,
|
||||
label: string,
|
||||
): TArgs {
|
||||
const call = mock.mock.calls[index];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} mock call ${index}`);
|
||||
}
|
||||
return call;
|
||||
}
|
||||
|
||||
function requireRequestInit(call: unknown[], label: string): RequestInit {
|
||||
const init = call[1];
|
||||
if (!init || typeof init !== "object") {
|
||||
throw new Error(`Expected ${label} request init`);
|
||||
}
|
||||
return init as RequestInit;
|
||||
}
|
||||
|
||||
function expectHeaders(init: RequestInit, expected: Record<string, string>): void {
|
||||
const headers = new Headers(init.headers);
|
||||
for (const [key, value] of Object.entries(expected)) {
|
||||
expect(headers.get(key)).toBe(value);
|
||||
}
|
||||
}
|
||||
|
||||
function parseRequestJsonBody(init: RequestInit): Record<string, unknown> {
|
||||
const requestBody = init.body;
|
||||
if (typeof requestBody !== "string") {
|
||||
throw new Error("Expected request body to be serialized JSON");
|
||||
}
|
||||
return JSON.parse(requestBody) as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireGenerationConfig(params: { generationConfig?: unknown }): Record<string, unknown> {
|
||||
const config = params.generationConfig;
|
||||
if (!config || typeof config !== "object") {
|
||||
throw new Error("Expected generationConfig");
|
||||
}
|
||||
return config as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireThinkingConfig(config: Record<string, unknown>): Record<string, unknown> {
|
||||
const thinkingConfig = config.thinkingConfig;
|
||||
if (!thinkingConfig || typeof thinkingConfig !== "object") {
|
||||
throw new Error("Expected thinkingConfig");
|
||||
}
|
||||
return thinkingConfig as Record<string, unknown>;
|
||||
}
|
||||
|
||||
describe("google transport stream", () => {
|
||||
beforeAll(async () => {
|
||||
({
|
||||
@@ -277,53 +226,57 @@ describe("google transport stream", () => {
|
||||
const result = await stream.result();
|
||||
|
||||
expect(buildGuardedModelFetchMock).toHaveBeenCalledWith(model);
|
||||
const guardedCall = requireMockCall(guardedFetchMock, 0, "guarded fetch");
|
||||
expect(guardedCall[0]).toBe(
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3.1-pro-preview:streamGenerateContent?alt=sse",
|
||||
expect.objectContaining({
|
||||
method: "POST",
|
||||
headers: expect.objectContaining({
|
||||
accept: "text/event-stream",
|
||||
"Content-Type": "application/json",
|
||||
"x-goog-api-key": "gemini-api-key",
|
||||
"X-Provider": "google",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const init = requireRequestInit(guardedCall, "guarded fetch");
|
||||
expect(init.method).toBe("POST");
|
||||
expectHeaders(init, {
|
||||
accept: "text/event-stream",
|
||||
"Content-Type": "application/json",
|
||||
"x-goog-api-key": "gemini-api-key",
|
||||
"X-Provider": "google",
|
||||
});
|
||||
|
||||
const payload = parseRequestJsonBody(init);
|
||||
const init = guardedFetchMock.mock.calls[0]?.[1] as RequestInit;
|
||||
const requestBody = init.body;
|
||||
if (typeof requestBody !== "string") {
|
||||
throw new Error("Expected Google transport request body to be serialized JSON");
|
||||
}
|
||||
const payload = JSON.parse(requestBody) as Record<string, unknown>;
|
||||
expect(payload.systemInstruction).toEqual({
|
||||
parts: [{ text: "Follow policy." }],
|
||||
});
|
||||
expect(payload.cachedContent).toBe("cachedContents/request-cache");
|
||||
expect((payload.generationConfig as { thinkingConfig?: unknown }).thinkingConfig).toEqual({
|
||||
includeThoughts: true,
|
||||
thinkingLevel: "HIGH",
|
||||
expect(payload.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true, thinkingLevel: "HIGH" },
|
||||
});
|
||||
expect(
|
||||
(payload.toolConfig as { functionCallingConfig?: unknown }).functionCallingConfig,
|
||||
).toEqual({
|
||||
mode: "AUTO",
|
||||
expect(payload.toolConfig).toMatchObject({
|
||||
functionCallingConfig: { mode: "AUTO" },
|
||||
});
|
||||
expect(result.api).toBe("google-generative-ai");
|
||||
expect(result.provider).toBe("google");
|
||||
expect(result.responseId).toBe("resp_1");
|
||||
expect(result.stopReason).toBe("toolUse");
|
||||
expect(result.usage.input).toBe(8);
|
||||
expect(result.usage.output).toBe(8);
|
||||
expect(result.usage.cacheRead).toBe(2);
|
||||
expect(result.usage.totalTokens).toBe(18);
|
||||
expect(result.content).toHaveLength(3);
|
||||
expect(result.content[0]).toEqual({
|
||||
type: "thinking",
|
||||
thinking: "draft",
|
||||
thinkingSignature: "sig_1",
|
||||
expect(result).toMatchObject({
|
||||
api: "google-generative-ai",
|
||||
provider: "google",
|
||||
responseId: "resp_1",
|
||||
stopReason: "toolUse",
|
||||
usage: {
|
||||
input: 8,
|
||||
output: 8,
|
||||
cacheRead: 2,
|
||||
totalTokens: 18,
|
||||
},
|
||||
content: [
|
||||
{ type: "thinking", thinking: "draft", thinkingSignature: "sig_1" },
|
||||
{ type: "text", text: "answer" },
|
||||
{
|
||||
type: "toolCall",
|
||||
name: "lookup",
|
||||
arguments: { q: "hello" },
|
||||
thoughtSignature: "call_sig_1",
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(result.content[1]?.type).toBe("text");
|
||||
expect(result.content[1]).toHaveProperty("text", "answer");
|
||||
expect(result.content[2]?.type).toBe("toolCall");
|
||||
expect(result.content[2]).toHaveProperty("name", "lookup");
|
||||
expect(result.content[2]).toHaveProperty("arguments", { q: "hello" });
|
||||
expect(result.content[2]).toHaveProperty("thoughtSignature", "call_sig_1");
|
||||
});
|
||||
|
||||
it("builds a lean Gemini 3 first-response retry payload", () => {
|
||||
@@ -400,7 +353,7 @@ describe("google transport stream", () => {
|
||||
expect(guardedFetchMock).toHaveBeenCalledTimes(2);
|
||||
const firstBody = JSON.parse(guardedFetchMock.mock.calls[0]?.[1]?.body as string);
|
||||
const retryBody = JSON.parse(guardedFetchMock.mock.calls[1]?.[1]?.body as string);
|
||||
expect(firstBody.generationConfig.thinkingConfig).toEqual({
|
||||
expect(firstBody.generationConfig.thinkingConfig).toMatchObject({
|
||||
includeThoughts: true,
|
||||
thinkingLevel: "HIGH",
|
||||
});
|
||||
@@ -481,13 +434,15 @@ describe("google transport stream", () => {
|
||||
);
|
||||
await stream.result();
|
||||
|
||||
const guardedCall = requireMockCall(guardedFetchMock, 0, "guarded fetch");
|
||||
expect(typeof guardedCall[0]).toBe("string");
|
||||
const init = requireRequestInit(guardedCall, "guarded fetch");
|
||||
expectHeaders(init, {
|
||||
Authorization: "Bearer oauth-token",
|
||||
"Content-Type": "application/json",
|
||||
});
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
expect.any(String),
|
||||
expect.objectContaining({
|
||||
headers: expect.objectContaining({
|
||||
Authorization: "Bearer oauth-token",
|
||||
"Content-Type": "application/json",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("refreshes authorized_user ADC before Google Vertex requests", async () => {
|
||||
@@ -539,25 +494,27 @@ describe("google transport stream", () => {
|
||||
);
|
||||
const result = await stream.result();
|
||||
|
||||
const tokenCall = requireMockCall(tokenFetchMock, 0, "token fetch");
|
||||
expect(tokenCall[0]).toBe("https://oauth2.googleapis.com/token");
|
||||
expect(requireRequestInit(tokenCall, "token fetch").method).toBe("POST");
|
||||
|
||||
const guardedCall = requireMockCall(guardedFetchMock, 0, "guarded fetch");
|
||||
expect(guardedCall[0]).toBe(
|
||||
"https://aiplatform.googleapis.com/v1/projects/vertex-project/locations/global/publishers/google/models/gemini-3.1-pro-preview:streamGenerateContent?alt=sse",
|
||||
expect(tokenFetchMock).toHaveBeenCalledWith(
|
||||
"https://oauth2.googleapis.com/token",
|
||||
expect.objectContaining({ method: "POST" }),
|
||||
);
|
||||
const guardedInit = requireRequestInit(guardedCall, "guarded fetch");
|
||||
expect(guardedInit.method).toBe("POST");
|
||||
expectHeaders(guardedInit, {
|
||||
Authorization: "Bearer ya29.vertex-token",
|
||||
"Content-Type": "application/json",
|
||||
accept: "text/event-stream",
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
"https://aiplatform.googleapis.com/v1/projects/vertex-project/locations/global/publishers/google/models/gemini-3.1-pro-preview:streamGenerateContent?alt=sse",
|
||||
expect.objectContaining({
|
||||
method: "POST",
|
||||
headers: expect.objectContaining({
|
||||
Authorization: "Bearer ya29.vertex-token",
|
||||
"Content-Type": "application/json",
|
||||
accept: "text/event-stream",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(result).toMatchObject({
|
||||
api: "google-vertex",
|
||||
provider: "google-vertex",
|
||||
stopReason: "stop",
|
||||
content: [{ type: "text", text: "ok" }],
|
||||
});
|
||||
expect(result.api).toBe("google-vertex");
|
||||
expect(result.provider).toBe("google-vertex");
|
||||
expect(result.stopReason).toBe("stop");
|
||||
expect(result.content).toEqual([{ type: "text", text: "ok" }]);
|
||||
});
|
||||
|
||||
it("refreshes authorized_user ADC from the Windows APPDATA fallback for Google Vertex requests", async () => {
|
||||
@@ -613,18 +570,25 @@ describe("google transport stream", () => {
|
||||
);
|
||||
await stream.result();
|
||||
|
||||
const tokenCall = requireMockCall(tokenFetchMock, 0, "token fetch");
|
||||
expect(tokenCall[0]).toBe("https://oauth2.googleapis.com/token");
|
||||
const tokenInit = requireRequestInit(tokenCall, "token fetch");
|
||||
expect(tokenInit.method).toBe("POST");
|
||||
expect(tokenInit.body).toBeInstanceOf(URLSearchParams);
|
||||
const requestBody = tokenInit.body as URLSearchParams;
|
||||
expect(tokenFetchMock).toHaveBeenCalledWith(
|
||||
"https://oauth2.googleapis.com/token",
|
||||
expect.objectContaining({
|
||||
body: expect.objectContaining({
|
||||
get: expect.any(Function),
|
||||
}),
|
||||
method: "POST",
|
||||
}),
|
||||
);
|
||||
const requestBody = tokenFetchMock.mock.calls[0]?.[1]?.body as URLSearchParams | undefined;
|
||||
expect(requestBody?.get("refresh_token")).toBe("appdata-refresh-token");
|
||||
const guardedCall = requireMockCall(guardedFetchMock, 0, "guarded fetch");
|
||||
expect(typeof guardedCall[0]).toBe("string");
|
||||
expectHeaders(requireRequestInit(guardedCall, "guarded fetch"), {
|
||||
Authorization: "Bearer ya29.appdata-token",
|
||||
});
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
expect.any(String),
|
||||
expect.objectContaining({
|
||||
headers: expect.objectContaining({
|
||||
Authorization: "Bearer ya29.appdata-token",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("coerces replayed malformed tool-call args to an object for Google payloads", () => {
|
||||
@@ -649,7 +613,7 @@ describe("google transport stream", () => {
|
||||
],
|
||||
} as never);
|
||||
|
||||
expect(params.contents[0]).toEqual({
|
||||
expect(params.contents[0]).toMatchObject({
|
||||
role: "model",
|
||||
parts: [{ functionCall: { name: "lookup", args: {} } }],
|
||||
});
|
||||
@@ -683,7 +647,7 @@ describe("google transport stream", () => {
|
||||
],
|
||||
} as never);
|
||||
|
||||
expect(params.contents[0]).toEqual({
|
||||
expect(params.contents[0]).toMatchObject({
|
||||
role: "model",
|
||||
parts: [
|
||||
{
|
||||
@@ -721,7 +685,7 @@ describe("google transport stream", () => {
|
||||
],
|
||||
} as never);
|
||||
|
||||
expect(params.contents[0]).toEqual({
|
||||
expect(params.contents[0]).toMatchObject({
|
||||
role: "model",
|
||||
parts: [
|
||||
{
|
||||
@@ -760,7 +724,7 @@ describe("google transport stream", () => {
|
||||
],
|
||||
} as never);
|
||||
|
||||
expect(params.contents[0]).toEqual({
|
||||
expect(params.contents[0]).toMatchObject({
|
||||
role: "model",
|
||||
parts: [{ functionCall: { name: "lookup", args: { q: "hello" } } }],
|
||||
});
|
||||
@@ -792,10 +756,12 @@ describe("google transport stream", () => {
|
||||
},
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
const thinkingConfig = requireThinkingConfig(generationConfig);
|
||||
expect(thinkingConfig.includeThoughts).toBe(true);
|
||||
expect(thinkingConfig).not.toHaveProperty("thinkingBudget");
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true },
|
||||
});
|
||||
expect(params.generationConfig).not.toMatchObject({
|
||||
thinkingConfig: { thinkingBudget: -1 },
|
||||
});
|
||||
});
|
||||
|
||||
it("omits disabled thinkingBudget=0 for Gemini 2.5 Pro direct payloads", () => {
|
||||
@@ -809,9 +775,10 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
expect(generationConfig.maxOutputTokens).toBe(128);
|
||||
expect(generationConfig).not.toHaveProperty("thinkingConfig");
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
maxOutputTokens: 128,
|
||||
});
|
||||
expect(params.generationConfig).not.toHaveProperty("thinkingConfig");
|
||||
});
|
||||
|
||||
it("strips explicit thinkingBudget=0 but preserves includeThoughts for Gemini 2.5 Pro", () => {
|
||||
@@ -828,10 +795,12 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
const thinkingConfig = requireThinkingConfig(generationConfig);
|
||||
expect(thinkingConfig.includeThoughts).toBe(true);
|
||||
expect(thinkingConfig).not.toHaveProperty("thinkingBudget");
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true },
|
||||
});
|
||||
expect(params.generationConfig).not.toMatchObject({
|
||||
thinkingConfig: { thinkingBudget: 0 },
|
||||
});
|
||||
});
|
||||
|
||||
it.each([
|
||||
@@ -851,11 +820,13 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
const thinkingConfig = requireThinkingConfig(generationConfig);
|
||||
expect(generationConfig.maxOutputTokens).toBe(128);
|
||||
expect(thinkingConfig.thinkingLevel).toBe(level);
|
||||
expect(thinkingConfig).not.toHaveProperty("thinkingBudget");
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
maxOutputTokens: 128,
|
||||
thinkingConfig: { thinkingLevel: level },
|
||||
});
|
||||
expect(params.generationConfig).not.toMatchObject({
|
||||
thinkingConfig: { thinkingBudget: 0 },
|
||||
});
|
||||
},
|
||||
);
|
||||
|
||||
@@ -873,13 +844,12 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
const thinkingConfig = requireThinkingConfig(generationConfig);
|
||||
expect(thinkingConfig).toEqual({
|
||||
includeThoughts: true,
|
||||
thinkingLevel: "MEDIUM",
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true, thinkingLevel: "MEDIUM" },
|
||||
});
|
||||
expect(params.generationConfig).not.toMatchObject({
|
||||
thinkingConfig: { thinkingBudget: 8192 },
|
||||
});
|
||||
expect(thinkingConfig).not.toHaveProperty("thinkingBudget");
|
||||
});
|
||||
|
||||
it("keeps adaptive Gemini 3 thinking on provider dynamic defaults", () => {
|
||||
@@ -893,11 +863,15 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
const thinkingConfig = requireThinkingConfig(generationConfig);
|
||||
expect(thinkingConfig.includeThoughts).toBe(true);
|
||||
expect(thinkingConfig).not.toHaveProperty("thinkingLevel");
|
||||
expect(thinkingConfig).not.toHaveProperty("thinkingBudget");
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true },
|
||||
});
|
||||
expect(params.generationConfig).not.toMatchObject({
|
||||
thinkingConfig: { thinkingLevel: expect.any(String) },
|
||||
});
|
||||
expect(params.generationConfig).not.toMatchObject({
|
||||
thinkingConfig: { thinkingBudget: expect.any(Number) },
|
||||
});
|
||||
});
|
||||
|
||||
it("maps adaptive Gemini 2.5 thinking to dynamic thinkingBudget", () => {
|
||||
@@ -911,10 +885,8 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
expect(requireThinkingConfig(generationConfig)).toEqual({
|
||||
includeThoughts: true,
|
||||
thinkingBudget: -1,
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true, thinkingBudget: -1 },
|
||||
});
|
||||
});
|
||||
|
||||
@@ -932,10 +904,8 @@ describe("google transport stream", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
expect(requireThinkingConfig(generationConfig)).toEqual({
|
||||
includeThoughts: true,
|
||||
thinkingLevel: "LOW",
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true, thinkingLevel: "LOW" },
|
||||
});
|
||||
});
|
||||
|
||||
@@ -1021,10 +991,8 @@ describe("google transport stream", () => {
|
||||
{ reasoning },
|
||||
);
|
||||
|
||||
const generationConfig = requireGenerationConfig(params);
|
||||
expect(requireThinkingConfig(generationConfig)).toEqual({
|
||||
includeThoughts: true,
|
||||
thinkingBudget: expectedBudget,
|
||||
expect(params.generationConfig).toMatchObject({
|
||||
thinkingConfig: { includeThoughts: true, thinkingBudget: expectedBudget },
|
||||
});
|
||||
});
|
||||
|
||||
@@ -1091,8 +1059,7 @@ describe("google transport stream", () => {
|
||||
"text_end",
|
||||
"done",
|
||||
]);
|
||||
expect(events[3]?.type).toBe("thinking_delta");
|
||||
expect(events[3]).toHaveProperty("delta", "");
|
||||
expect(events[3]).toMatchObject({ type: "thinking_delta", delta: "" });
|
||||
});
|
||||
|
||||
it("starts a thinking block for thoughtSignature-only parts that arrive before any text", async () => {
|
||||
|
||||
@@ -127,95 +127,6 @@ function createReactionHarness(params?: {
|
||||
});
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
expect(value, label).toBeTypeOf("object");
|
||||
expect(value, label).not.toBeNull();
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): Array<unknown> {
|
||||
expect(Array.isArray(value), label).toBe(true);
|
||||
return value as Array<unknown>;
|
||||
}
|
||||
|
||||
function mockCalls(mock: unknown, label: string): Array<Array<unknown>> {
|
||||
const mockState = (mock as { mock?: { calls?: Array<Array<unknown>> } }).mock;
|
||||
expect(mockState, `${label}.mock`).toBeDefined();
|
||||
expect(Array.isArray(mockState?.calls), `${label}.mock.calls`).toBe(true);
|
||||
return mockState?.calls ?? [];
|
||||
}
|
||||
|
||||
function callArg(mock: unknown, callIndex: number, argIndex: number, label: string) {
|
||||
const call = mockCalls(mock, label).at(callIndex);
|
||||
expect(call, label).toBeDefined();
|
||||
return call?.[argIndex];
|
||||
}
|
||||
|
||||
function lastCallArg(mock: unknown, argIndex: number, label: string) {
|
||||
const calls = mockCalls(mock, label);
|
||||
return callArg(mock, calls.length - 1, argIndex, label);
|
||||
}
|
||||
|
||||
function expectMockCallWithFields(mock: unknown, fields: Record<string, unknown>) {
|
||||
const matched = mockCalls(mock, "mock calls").some(([value]) => {
|
||||
if (!value || typeof value !== "object") {
|
||||
return false;
|
||||
}
|
||||
const record = value as Record<string, unknown>;
|
||||
return Object.entries(fields).every(([key, expected]) => Object.is(record[key], expected));
|
||||
});
|
||||
expect(matched).toBe(true);
|
||||
}
|
||||
|
||||
function expectNoticeSent(mock: unknown) {
|
||||
const message = requireRecord(callArg(mock, 0, 1, "notice content"), "notice content");
|
||||
expect(message.msgtype).toBe("m.notice");
|
||||
expect(String(message.body)).toContain("channels.matrix.dm.sessionScope");
|
||||
}
|
||||
|
||||
function findMockCall(mock: unknown, label: string, predicate: (call: Array<unknown>) => boolean) {
|
||||
const call = mockCalls(mock, label).find(predicate);
|
||||
expect(call, label).toBeDefined();
|
||||
return call as Array<unknown>;
|
||||
}
|
||||
|
||||
function expectFinalizedPreviewEdit(eventId: string, text: string) {
|
||||
const call = findMockCall(
|
||||
editMessageMatrixMock,
|
||||
`edit call for ${eventId}`,
|
||||
([room, editedEventId, body]) =>
|
||||
room === "!room:example.org" && editedEventId === eventId && body === text,
|
||||
);
|
||||
const options = requireRecord(call[3], "edit options");
|
||||
expect(options.extraContent).toEqual({ [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true });
|
||||
}
|
||||
|
||||
function expectEditLiveFlag(eventId: string, text: string, expected: boolean | undefined) {
|
||||
const call = findMockCall(
|
||||
editMessageMatrixMock,
|
||||
`edit live flag call for ${eventId}`,
|
||||
([room, editedEventId, body]) =>
|
||||
room === "!room:example.org" && editedEventId === eventId && body === text,
|
||||
);
|
||||
const options = requireRecord(call[3], "edit options");
|
||||
if (expected === undefined) {
|
||||
expect(Object.hasOwn(options, "live")).toBe(false);
|
||||
} else {
|
||||
expect(options.live).toBe(expected);
|
||||
}
|
||||
}
|
||||
|
||||
function expectDeliveredMediaReply() {
|
||||
const payload = requireRecord(
|
||||
lastCallArg(deliverMatrixRepliesMock, 0, "deliver replies payload"),
|
||||
"deliver replies payload",
|
||||
);
|
||||
const replies = requireArray(payload.replies, "deliver replies");
|
||||
const reply = requireRecord(replies[0], "media reply");
|
||||
expect(reply.mediaUrl).toBe("https://example.com/image.png");
|
||||
expect(reply.text).toBeUndefined();
|
||||
}
|
||||
|
||||
describe("matrix monitor handler pairing account scope", () => {
|
||||
it("caches account-scoped allowFrom store reads on hot path", async () => {
|
||||
const readAllowFromStore = vi.fn(async () => [] as string[]);
|
||||
@@ -303,16 +214,18 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const inbound = requireRecord(
|
||||
callArg(recordInboundSession, 0, 0, "record inbound session"),
|
||||
"record inbound session",
|
||||
expect(recordInboundSession).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
updateLastRoute: expect.objectContaining({
|
||||
channel: "matrix",
|
||||
to: "room:!dm:example.org",
|
||||
mainDmOwnerPin: expect.objectContaining({
|
||||
ownerRecipient: "@owner:example.org",
|
||||
senderRecipient: "@owner:example.org",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const route = requireRecord(inbound.updateLastRoute, "last route update");
|
||||
expect(route.channel).toBe("matrix");
|
||||
expect(route.to).toBe("room:!dm:example.org");
|
||||
const ownerPin = requireRecord(route.mainDmOwnerPin, "main DM owner pin");
|
||||
expect(ownerPin.ownerRecipient).toBe("@owner:example.org");
|
||||
expect(ownerPin.senderRecipient).toBe("@owner:example.org");
|
||||
});
|
||||
|
||||
it("uses live dmScope when deciding whether to pin main DM route updates", async () => {
|
||||
@@ -350,12 +263,13 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const inbound = requireRecord(
|
||||
callArg(recordInboundSession, 0, 0, "record inbound session"),
|
||||
"record inbound session",
|
||||
expect(recordInboundSession).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
updateLastRoute: expect.objectContaining({
|
||||
mainDmOwnerPin: undefined,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const route = requireRecord(inbound.updateLastRoute, "last route update");
|
||||
expect(route.mainDmOwnerPin).toBeUndefined();
|
||||
});
|
||||
|
||||
it("sends pairing reminders for pending requests with cooldown", async () => {
|
||||
@@ -456,7 +370,12 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expectMockCallWithFields(resolveAgentRoute, { channel: "matrix", accountId: "ops" });
|
||||
expect(resolveAgentRoute).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "matrix",
|
||||
accountId: "ops",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("does not enqueue delivered text messages into system events", async () => {
|
||||
@@ -735,14 +654,14 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
);
|
||||
|
||||
expect(hasControlCommand).toHaveBeenCalledWith("/new", expect.anything());
|
||||
const context = requireRecord(
|
||||
callArg(finalizeInboundContext, 0, 0, "finalized context"),
|
||||
"finalized context",
|
||||
expect(finalizeInboundContext).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
RawBody: "@bot:example.org /new",
|
||||
CommandBody: "/new",
|
||||
BodyForAgent: "@bot:example.org /new",
|
||||
BodyForCommands: "/new",
|
||||
}),
|
||||
);
|
||||
expect(context.RawBody).toBe("@bot:example.org /new");
|
||||
expect(context.CommandBody).toBe("/new");
|
||||
expect(context.BodyForAgent).toBe("@bot:example.org /new");
|
||||
expect(context.BodyForCommands).toBe("/new");
|
||||
expect(recordInboundSession).toHaveBeenCalled();
|
||||
});
|
||||
|
||||
@@ -979,13 +898,17 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const context = requireRecord(
|
||||
callArg(finalizeInboundContext, 0, 0, "finalized context"),
|
||||
"finalized context",
|
||||
expect(finalizeInboundContext).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
MessageThreadId: "$root",
|
||||
ThreadStarterBody: "Matrix thread root $root from Alice:\nRoot topic",
|
||||
}),
|
||||
);
|
||||
expect(recordInboundSession).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: "agent:ops:main:thread:$root",
|
||||
}),
|
||||
);
|
||||
expect(context.MessageThreadId).toBe("$root");
|
||||
expect(context.ThreadStarterBody).toBe("Matrix thread root $root from Alice:\nRoot topic");
|
||||
expectMockCallWithFields(recordInboundSession, { sessionKey: "agent:ops:main:thread:$root" });
|
||||
});
|
||||
|
||||
it("keeps threaded DMs flat when dm threadReplies is off", async () => {
|
||||
@@ -1021,14 +944,18 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const context = requireRecord(
|
||||
callArg(finalizeInboundContext, 0, 0, "finalized context"),
|
||||
"finalized context",
|
||||
expect(finalizeInboundContext).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
MessageThreadId: undefined,
|
||||
ReplyToId: "$root",
|
||||
ThreadStarterBody: "Matrix thread root $root from Alice:\nRoot topic",
|
||||
}),
|
||||
);
|
||||
expect(recordInboundSession).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: "agent:ops:main",
|
||||
}),
|
||||
);
|
||||
expect(context.MessageThreadId).toBeUndefined();
|
||||
expect(context.ReplyToId).toBe("$root");
|
||||
expect(context.ThreadStarterBody).toBe("Matrix thread root $root from Alice:\nRoot topic");
|
||||
expectMockCallWithFields(recordInboundSession, { sessionKey: "agent:ops:main" });
|
||||
});
|
||||
|
||||
it("posts a one-time notice when another Matrix DM room already owns the shared DM session", async () => {
|
||||
@@ -1060,8 +987,13 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expect(sendNotice).toHaveBeenCalledWith("!dm:example.org", expect.anything());
|
||||
expectNoticeSent(sendNotice);
|
||||
expect(sendNotice).toHaveBeenCalledWith(
|
||||
"!dm:example.org",
|
||||
expect.objectContaining({
|
||||
msgtype: "m.notice",
|
||||
body: expect.stringContaining("channels.matrix.dm.sessionScope"),
|
||||
}),
|
||||
);
|
||||
|
||||
await handler(
|
||||
"!dm:example.org",
|
||||
@@ -1114,8 +1046,13 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expect(sendNotice).toHaveBeenCalledWith("!dm:example.org", expect.anything());
|
||||
expectNoticeSent(sendNotice);
|
||||
expect(sendNotice).toHaveBeenCalledWith(
|
||||
"!dm:example.org",
|
||||
expect.objectContaining({
|
||||
msgtype: "m.notice",
|
||||
body: expect.stringContaining("channels.matrix.dm.sessionScope"),
|
||||
}),
|
||||
);
|
||||
} finally {
|
||||
fs.rmSync(tempDir, { recursive: true, force: true });
|
||||
}
|
||||
@@ -1166,8 +1103,13 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expect(sendNotice).toHaveBeenCalledWith("!dm:example.org", expect.anything());
|
||||
expectNoticeSent(sendNotice);
|
||||
expect(sendNotice).toHaveBeenCalledWith(
|
||||
"!dm:example.org",
|
||||
expect.objectContaining({
|
||||
msgtype: "m.notice",
|
||||
body: expect.stringContaining("channels.matrix.dm.sessionScope"),
|
||||
}),
|
||||
);
|
||||
} finally {
|
||||
fs.rmSync(tempDir, { recursive: true, force: true });
|
||||
}
|
||||
@@ -1208,8 +1150,13 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expect(sendNotice).toHaveBeenCalledWith("!dm:example.org", expect.anything());
|
||||
expectNoticeSent(sendNotice);
|
||||
expect(sendNotice).toHaveBeenCalledWith(
|
||||
"!dm:example.org",
|
||||
expect.objectContaining({
|
||||
msgtype: "m.notice",
|
||||
body: expect.stringContaining("channels.matrix.dm.sessionScope"),
|
||||
}),
|
||||
);
|
||||
} finally {
|
||||
fs.rmSync(tempDir, { recursive: true, force: true });
|
||||
}
|
||||
@@ -1289,9 +1236,11 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
);
|
||||
|
||||
expect(sendNotice).not.toHaveBeenCalled();
|
||||
expectMockCallWithFields(recordInboundSession, {
|
||||
sessionKey: "agent:ops:matrix:channel:!dm:example.org",
|
||||
});
|
||||
expect(recordInboundSession).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: "agent:ops:matrix:channel:!dm:example.org",
|
||||
}),
|
||||
);
|
||||
} finally {
|
||||
fs.rmSync(tempDir, { recursive: true, force: true });
|
||||
}
|
||||
@@ -1390,13 +1339,14 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
const finalized = requireRecord(
|
||||
lastCallArg(finalizeInboundContext, 0, "finalized context"),
|
||||
"finalized context",
|
||||
const finalized = vi.mocked(finalizeInboundContext).mock.calls.at(-1)?.[0];
|
||||
expect(finalized).toEqual(
|
||||
expect.objectContaining({
|
||||
GroupChannel: "!room:example.org",
|
||||
GroupSubject: "Ops Room",
|
||||
GroupId: "!room:example.org",
|
||||
}),
|
||||
);
|
||||
expect(finalized.GroupChannel).toBe("!room:example.org");
|
||||
expect(finalized.GroupSubject).toBe("Ops Room");
|
||||
expect(finalized.GroupId).toBe("!room:example.org");
|
||||
});
|
||||
|
||||
it("routes bound Matrix threads to the target session key", async () => {
|
||||
@@ -1454,7 +1404,11 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expectMockCallWithFields(recordInboundSession, { sessionKey: "agent:bound:session-1" });
|
||||
expect(recordInboundSession).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: "agent:bound:session-1",
|
||||
}),
|
||||
);
|
||||
expect(touch).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
@@ -1659,7 +1613,12 @@ describe("matrix monitor handler pairing account scope", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
expectMockCallWithFields(resolveAgentRoute, { channel: "matrix", accountId: "ops" });
|
||||
expect(resolveAgentRoute).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "matrix",
|
||||
accountId: "ops",
|
||||
}),
|
||||
);
|
||||
expect(enqueueSystemEvent).toHaveBeenCalledWith(
|
||||
"Matrix reaction added: 👍 by sender on msg $msg1",
|
||||
{
|
||||
@@ -2186,12 +2145,12 @@ describe("matrix monitor handler live allowlist reload", () => {
|
||||
body: "hello again",
|
||||
});
|
||||
|
||||
const liveAllowlistRequest = requireRecord(
|
||||
lastCallArg(resolveLiveUserAllowlist, 0, "live allowlist request"),
|
||||
"live allowlist request",
|
||||
expect(resolveLiveUserAllowlist).toHaveBeenLastCalledWith(
|
||||
expect.objectContaining({
|
||||
accountId: "ops",
|
||||
entries: ["Alice"],
|
||||
}),
|
||||
);
|
||||
expect(liveAllowlistRequest.accountId).toBe("ops");
|
||||
expect(liveAllowlistRequest.entries).toEqual(["Alice"]);
|
||||
expect(dispatchReplyFromConfig).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
@@ -2765,7 +2724,14 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
await deliver({ text: "Single block" }, { kind: "final" });
|
||||
|
||||
expect(sendSingleTextMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
expectFinalizedPreviewEdit("$draft1", "Single block");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"Single block",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
expect(deliverMatrixRepliesMock).not.toHaveBeenCalled();
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
await finish();
|
||||
@@ -2787,7 +2753,14 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
|
||||
await deliver({ text: "Done" }, { kind: "final" });
|
||||
|
||||
expectFinalizedPreviewEdit("$draft1", "Done");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"Done",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
expect(deliverMatrixRepliesMock).not.toHaveBeenCalled();
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
await finish();
|
||||
@@ -2895,19 +2868,26 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
expect(sendSingleTextMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
const draftOptions = requireRecord(
|
||||
callArg(sendSingleTextMessageMatrixMock, 0, 2, "draft options"),
|
||||
"draft options",
|
||||
expect(sendSingleTextMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"Single block",
|
||||
expect.not.objectContaining({
|
||||
msgtype: "m.notice",
|
||||
includeMentions: false,
|
||||
}),
|
||||
);
|
||||
expect(draftOptions.msgtype).not.toBe("m.notice");
|
||||
expect(draftOptions.includeMentions).not.toBe(false);
|
||||
|
||||
await deliver({ text: "Single block" }, { kind: "final" });
|
||||
|
||||
// MSC4357: even when text is unchanged, a finalize edit is sent to clear
|
||||
// the live marker so supporting clients stop the streaming animation.
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
expectEditLiveFlag("$draft1", "Single block", false);
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"Single block",
|
||||
expect.objectContaining({ live: false }),
|
||||
);
|
||||
expect(deliverMatrixRepliesMock).not.toHaveBeenCalled();
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
await finish();
|
||||
@@ -2928,7 +2908,12 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
await deliver({ text: "Single block" }, { kind: "final" });
|
||||
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
expectEditLiveFlag("$draft1", "Single block", undefined);
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"Single block",
|
||||
expect.not.objectContaining({ live: false }),
|
||||
);
|
||||
expect(deliverMatrixRepliesMock).not.toHaveBeenCalled();
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
await finish();
|
||||
@@ -2946,7 +2931,14 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
deliverMatrixRepliesMock.mockClear();
|
||||
await deliver({ text: "Block one" }, { kind: "block" });
|
||||
|
||||
expectFinalizedPreviewEdit("$draft1", "Block one");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"Block one",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
expect(deliverMatrixRepliesMock).not.toHaveBeenCalled();
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
|
||||
@@ -2962,7 +2954,14 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
|
||||
await deliver({ text: "Block two" }, { kind: "final" });
|
||||
|
||||
expectFinalizedPreviewEdit("$draft2", "Block two");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft2",
|
||||
"Block two",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
expect(deliverMatrixRepliesMock).not.toHaveBeenCalled();
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
await finish();
|
||||
@@ -3260,7 +3259,14 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
expect(sendSingleTextMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
expect(sendSingleTextMessageMatrixMock.mock.calls[0]?.[1]).toBe("Beta");
|
||||
expectFinalizedPreviewEdit("$draft1", "Alpha");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"Alpha",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
|
||||
sendSingleTextMessageMatrixMock.mockClear();
|
||||
editMessageMatrixMock.mockClear();
|
||||
@@ -3274,7 +3280,14 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
expect(sendSingleTextMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
expect(sendSingleTextMessageMatrixMock.mock.calls[0]?.[1]).toBe("Gamma");
|
||||
expectFinalizedPreviewEdit("$draft2", "Beta");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft2",
|
||||
"Beta",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
|
||||
await finish();
|
||||
});
|
||||
@@ -3552,9 +3565,23 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
);
|
||||
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledTimes(1);
|
||||
expectEditLiveFlag("$draft1", "@room screenshot ready", false);
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"@room screenshot ready",
|
||||
expect.objectContaining({ live: false }),
|
||||
);
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
expectDeliveredMediaReply();
|
||||
expect(deliverMatrixRepliesMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
replies: [
|
||||
expect.objectContaining({
|
||||
mediaUrl: "https://example.com/image.png",
|
||||
text: undefined,
|
||||
}),
|
||||
],
|
||||
}),
|
||||
);
|
||||
await finish();
|
||||
});
|
||||
|
||||
@@ -3576,9 +3603,25 @@ describe("matrix monitor handler draft streaming", () => {
|
||||
{ kind: "final" },
|
||||
);
|
||||
|
||||
expectFinalizedPreviewEdit("$draft1", "@room screenshot ready");
|
||||
expect(editMessageMatrixMock).toHaveBeenCalledWith(
|
||||
"!room:example.org",
|
||||
"$draft1",
|
||||
"@room screenshot ready",
|
||||
expect.objectContaining({
|
||||
extraContent: { [MATRIX_OPENCLAW_FINALIZED_PREVIEW_KEY]: true },
|
||||
}),
|
||||
);
|
||||
expect(redactEventMock).not.toHaveBeenCalled();
|
||||
expectDeliveredMediaReply();
|
||||
expect(deliverMatrixRepliesMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
replies: [
|
||||
expect.objectContaining({
|
||||
mediaUrl: "https://example.com/image.png",
|
||||
text: undefined,
|
||||
}),
|
||||
],
|
||||
}),
|
||||
);
|
||||
await finish();
|
||||
});
|
||||
|
||||
|
||||
@@ -28,46 +28,6 @@ type GuardedFetchCall = {
|
||||
auditContext?: string;
|
||||
};
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
expect(value, label).toBeTypeOf("object");
|
||||
expect(value, label).not.toBeNull();
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireHeaders(value: unknown): Record<string, string> {
|
||||
return requireRecord(value, "request headers") as Record<string, string>;
|
||||
}
|
||||
|
||||
function expectToolCallContent(
|
||||
value: unknown,
|
||||
expected: { name: string; arguments: Record<string, unknown> },
|
||||
) {
|
||||
const content = requireRecord(value, "tool call content");
|
||||
expect(content.type).toBe("toolCall");
|
||||
expect(content.name).toBe(expected.name);
|
||||
expect(content.arguments).toEqual(expected.arguments);
|
||||
}
|
||||
|
||||
function expectIteratorEvent(
|
||||
value: unknown,
|
||||
expected: { type?: string; delta?: string; content?: string; done: boolean },
|
||||
) {
|
||||
const result = requireRecord(value, "iterator result");
|
||||
expect(result.done).toBe(expected.done);
|
||||
if (expected.type !== undefined) {
|
||||
const event = requireRecord(result.value, "iterator result value");
|
||||
expect(event.type).toBe(expected.type);
|
||||
if (expected.delta !== undefined) {
|
||||
expect(event.delta).toBe(expected.delta);
|
||||
}
|
||||
if (expected.content !== undefined) {
|
||||
expect(event.content).toBe(expected.content);
|
||||
}
|
||||
} else {
|
||||
expect(result.value).toBeUndefined();
|
||||
}
|
||||
}
|
||||
|
||||
afterEach(() => {
|
||||
fetchWithSsrFGuardMock.mockReset();
|
||||
});
|
||||
@@ -89,29 +49,38 @@ describe("buildOllamaChatRequest", () => {
|
||||
});
|
||||
|
||||
it("strips the ollama/ prefix from chat model ids", () => {
|
||||
const request = buildOllamaChatRequest({
|
||||
modelId: "ollama/qwen3:14b-q8_0",
|
||||
messages: [{ role: "user", content: "hello" }],
|
||||
expect(
|
||||
buildOllamaChatRequest({
|
||||
modelId: "ollama/qwen3:14b-q8_0",
|
||||
messages: [{ role: "user", content: "hello" }],
|
||||
}),
|
||||
).toMatchObject({
|
||||
model: "qwen3:14b-q8_0",
|
||||
});
|
||||
expect(request.model).toBe("qwen3:14b-q8_0");
|
||||
});
|
||||
|
||||
it("strips the active custom provider prefix from chat model ids", () => {
|
||||
const request = buildOllamaChatRequest({
|
||||
modelId: "ollama-spark/qwen3:32b",
|
||||
providerId: "ollama-spark",
|
||||
messages: [{ role: "user", content: "hello" }],
|
||||
expect(
|
||||
buildOllamaChatRequest({
|
||||
modelId: "ollama-spark/qwen3:32b",
|
||||
providerId: "ollama-spark",
|
||||
messages: [{ role: "user", content: "hello" }],
|
||||
}),
|
||||
).toMatchObject({
|
||||
model: "qwen3:32b",
|
||||
});
|
||||
expect(request.model).toBe("qwen3:32b");
|
||||
});
|
||||
|
||||
it("keeps unrelated slash-containing Ollama model ids intact", () => {
|
||||
const request = buildOllamaChatRequest({
|
||||
modelId: "library/qwen3:32b",
|
||||
providerId: "ollama-spark",
|
||||
messages: [{ role: "user", content: "hello" }],
|
||||
expect(
|
||||
buildOllamaChatRequest({
|
||||
modelId: "library/qwen3:32b",
|
||||
providerId: "ollama-spark",
|
||||
messages: [{ role: "user", content: "hello" }],
|
||||
}),
|
||||
).toMatchObject({
|
||||
model: "library/qwen3:32b",
|
||||
});
|
||||
expect(request.model).toBe("library/qwen3:32b");
|
||||
});
|
||||
});
|
||||
|
||||
@@ -149,9 +118,10 @@ describe("createConfiguredOllamaCompatStreamWrapper", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const payload = requireRecord(patchedPayload, "patched payload");
|
||||
expect(payload.thinking).toEqual({ type: "enabled" });
|
||||
expect(payload.options).toEqual({ num_ctx: 65536 });
|
||||
expect(patchedPayload).toMatchObject({
|
||||
thinking: { type: "enabled" },
|
||||
options: { num_ctx: 65536 },
|
||||
});
|
||||
});
|
||||
|
||||
it("falls back to contextWindow when configured num_ctx is invalid", async () => {
|
||||
@@ -185,8 +155,9 @@ describe("createConfiguredOllamaCompatStreamWrapper", () => {
|
||||
} as never,
|
||||
);
|
||||
|
||||
const payload = requireRecord(patchedPayload, "patched payload");
|
||||
expect(payload.options).toEqual({ num_ctx: 131072 });
|
||||
expect(patchedPayload).toMatchObject({
|
||||
options: { num_ctx: 131072 },
|
||||
});
|
||||
});
|
||||
|
||||
it("forwards think=false on native Ollama chat requests when thinking is off", async () => {
|
||||
@@ -237,7 +208,7 @@ describe("createConfiguredOllamaCompatStreamWrapper", () => {
|
||||
};
|
||||
expect(requestBody.think).toBe(false);
|
||||
expect(requestBody.options?.think).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBe(131072);
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -339,7 +310,7 @@ describe("createConfiguredOllamaCompatStreamWrapper", () => {
|
||||
};
|
||||
expect(requestBody.think).toBe("low");
|
||||
expect(requestBody.options?.think).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBe(131072);
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -434,7 +405,7 @@ describe("createConfiguredOllamaCompatStreamWrapper", () => {
|
||||
};
|
||||
expect(requestBody.think).toBe("high");
|
||||
expect(requestBody.options?.think).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBe(131072);
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -888,9 +859,14 @@ describe("buildAssistantMessage", () => {
|
||||
done: true,
|
||||
};
|
||||
const result = buildAssistantMessage(response, modelInfo);
|
||||
expect(result.content).toHaveLength(2);
|
||||
expectToolCallContent(result.content[0], { name: "exec", arguments: { command: "pwd" } });
|
||||
expectToolCallContent(result.content[1], { name: "read", arguments: { path: "README.md" } });
|
||||
expect(result.content).toEqual([
|
||||
expect.objectContaining({ type: "toolCall", name: "exec", arguments: { command: "pwd" } }),
|
||||
expect.objectContaining({
|
||||
type: "toolCall",
|
||||
name: "read",
|
||||
arguments: { path: "README.md" },
|
||||
}),
|
||||
]);
|
||||
});
|
||||
|
||||
it("preserves exact allowlisted tool-prefix names in Ollama responses", () => {
|
||||
@@ -911,13 +887,19 @@ describe("buildAssistantMessage", () => {
|
||||
const result = buildAssistantMessage(response, modelInfo, undefined, {
|
||||
availableToolNames: new Set(["tool_a", "tools_invoke_test", "function-run"]),
|
||||
});
|
||||
expect(result.content).toHaveLength(3);
|
||||
expectToolCallContent(result.content[0], { name: "tool_a", arguments: { value: 1 } });
|
||||
expectToolCallContent(result.content[1], {
|
||||
name: "tools_invoke_test",
|
||||
arguments: { value: 2 },
|
||||
});
|
||||
expectToolCallContent(result.content[2], { name: "function-run", arguments: { value: 3 } });
|
||||
expect(result.content).toEqual([
|
||||
expect.objectContaining({ type: "toolCall", name: "tool_a", arguments: { value: 1 } }),
|
||||
expect.objectContaining({
|
||||
type: "toolCall",
|
||||
name: "tools_invoke_test",
|
||||
arguments: { value: 2 },
|
||||
}),
|
||||
expect.objectContaining({
|
||||
type: "toolCall",
|
||||
name: "function-run",
|
||||
arguments: { value: 3 },
|
||||
}),
|
||||
]);
|
||||
});
|
||||
|
||||
it("keeps non-prefixed Ollama response tool names intact", () => {
|
||||
@@ -937,11 +919,12 @@ describe("buildAssistantMessage", () => {
|
||||
done: true,
|
||||
};
|
||||
const result = buildAssistantMessage(response, modelInfo);
|
||||
expect(result.content).toHaveLength(4);
|
||||
expectToolCallContent(result.content[0], { name: "functionshell", arguments: {} });
|
||||
expectToolCallContent(result.content[1], { name: "tooling", arguments: {} });
|
||||
expectToolCallContent(result.content[2], { name: "tools", arguments: {} });
|
||||
expectToolCallContent(result.content[3], { name: "tool_a", arguments: {} });
|
||||
expect(result.content).toEqual([
|
||||
expect.objectContaining({ type: "toolCall", name: "functionshell", arguments: {} }),
|
||||
expect.objectContaining({ type: "toolCall", name: "tooling", arguments: {} }),
|
||||
expect.objectContaining({ type: "toolCall", name: "tools", arguments: {} }),
|
||||
expect.objectContaining({ type: "toolCall", name: "tool_a", arguments: {} }),
|
||||
]);
|
||||
});
|
||||
|
||||
it("parses stringified tool call arguments from Ollama responses", () => {
|
||||
@@ -956,7 +939,8 @@ describe("buildAssistantMessage", () => {
|
||||
done: true,
|
||||
};
|
||||
const result = buildAssistantMessage(response, modelInfo);
|
||||
expectToolCallContent(result.content[0], {
|
||||
expect(result.content[0]).toMatchObject({
|
||||
type: "toolCall",
|
||||
name: "bash",
|
||||
arguments: { command: "ls", path: "/tmp" },
|
||||
});
|
||||
@@ -981,7 +965,8 @@ describe("buildAssistantMessage", () => {
|
||||
done: true,
|
||||
};
|
||||
const result = buildAssistantMessage(response, modelInfo);
|
||||
expectToolCallContent(result.content[0], {
|
||||
expect(result.content[0]).toMatchObject({
|
||||
type: "toolCall",
|
||||
name: "send",
|
||||
arguments: {
|
||||
target: "9223372036854775807",
|
||||
@@ -1002,7 +987,11 @@ describe("buildAssistantMessage", () => {
|
||||
done: true,
|
||||
};
|
||||
const result = buildAssistantMessage(response, modelInfo);
|
||||
expectToolCallContent(result.content[0], { name: "bash", arguments: {} });
|
||||
expect(result.content[0]).toMatchObject({
|
||||
type: "toolCall",
|
||||
name: "bash",
|
||||
arguments: {},
|
||||
});
|
||||
});
|
||||
|
||||
it("sets all costs to zero for local models", () => {
|
||||
@@ -1287,15 +1276,12 @@ describe("createOllamaStreamFn streaming events", () => {
|
||||
|
||||
// text_delta events carry incremental deltas
|
||||
const deltas = events.filter((e) => e.type === "text_delta");
|
||||
expect(deltas[0]?.contentIndex).toBe(0);
|
||||
expect(deltas[0]?.delta).toBe("Hello");
|
||||
expect(deltas[1]?.contentIndex).toBe(0);
|
||||
expect(deltas[1]?.delta).toBe(" world");
|
||||
expect(deltas[0]).toMatchObject({ contentIndex: 0, delta: "Hello" });
|
||||
expect(deltas[1]).toMatchObject({ contentIndex: 0, delta: " world" });
|
||||
|
||||
// text_end carries the full accumulated content
|
||||
const textEnd = events.find((e) => e.type === "text_end");
|
||||
expect(textEnd?.contentIndex).toBe(0);
|
||||
expect(textEnd?.content).toBe("Hello world");
|
||||
expect(textEnd).toMatchObject({ contentIndex: 0, content: "Hello world" });
|
||||
|
||||
// start/text_start carry empty partials (before any content accumulates)
|
||||
const startEvent = events.find((e) => e.type === "start");
|
||||
@@ -1437,11 +1423,10 @@ describe("createOllamaStreamFn streaming events", () => {
|
||||
expect(startEvent).not.toBe("timeout");
|
||||
expect(textStartEvent).not.toBe("timeout");
|
||||
expect(textDeltaEvent).not.toBe("timeout");
|
||||
expectIteratorEvent(startEvent, { type: "start", done: false });
|
||||
expectIteratorEvent(textStartEvent, { type: "text_start", done: false });
|
||||
expectIteratorEvent(textDeltaEvent, {
|
||||
type: "text_delta",
|
||||
delta: "Let me check.",
|
||||
expect(startEvent).toMatchObject({ value: { type: "start" }, done: false });
|
||||
expect(textStartEvent).toMatchObject({ value: { type: "text_start" }, done: false });
|
||||
expect(textDeltaEvent).toMatchObject({
|
||||
value: { type: "text_delta", delta: "Let me check." },
|
||||
done: false,
|
||||
});
|
||||
|
||||
@@ -1451,18 +1436,17 @@ describe("createOllamaStreamFn streaming events", () => {
|
||||
|
||||
const textEndEvent = await nextEventWithin(iterator);
|
||||
expect(textEndEvent).not.toBe("timeout");
|
||||
expectIteratorEvent(textEndEvent, {
|
||||
type: "text_end",
|
||||
content: "Let me check.",
|
||||
expect(textEndEvent).toMatchObject({
|
||||
value: {
|
||||
type: "text_end",
|
||||
contentIndex: 0,
|
||||
content: "Let me check.",
|
||||
partial: {
|
||||
content: [{ type: "text", text: "Let me check." }],
|
||||
},
|
||||
},
|
||||
done: false,
|
||||
});
|
||||
if (textEndEvent !== "timeout") {
|
||||
const textEndValue = requireRecord(textEndEvent.value, "text_end value");
|
||||
expect(textEndValue.contentIndex).toBe(0);
|
||||
expect(requireRecord(textEndValue.partial, "text_end partial").content).toEqual([
|
||||
{ type: "text", text: "Let me check." },
|
||||
]);
|
||||
}
|
||||
|
||||
controlledFetch.pushLine(
|
||||
'{"model":"m","created_at":"t","message":{"role":"assistant","content":""},"done":true,"prompt_eval_count":10,"eval_count":5}',
|
||||
@@ -1472,14 +1456,16 @@ describe("createOllamaStreamFn streaming events", () => {
|
||||
const doneEvent = await nextEventWithin(iterator);
|
||||
expect(doneEvent).not.toBe("timeout");
|
||||
if (doneEvent !== "timeout" && doneEvent.done === false) {
|
||||
expectIteratorEvent(doneEvent, { type: "done", done: false });
|
||||
expect(requireRecord(doneEvent.value, "done value").reason).toBe("toolUse");
|
||||
expect(doneEvent).toMatchObject({
|
||||
value: { type: "done", reason: "toolUse" },
|
||||
done: false,
|
||||
});
|
||||
|
||||
const streamEnd = await nextEventWithin(iterator);
|
||||
expect(streamEnd).not.toBe("timeout");
|
||||
expectIteratorEvent(streamEnd, { done: true });
|
||||
expect(streamEnd).toMatchObject({ value: undefined, done: true });
|
||||
} else {
|
||||
expectIteratorEvent(doneEvent, { done: true });
|
||||
expect(doneEvent).toMatchObject({ value: undefined, done: true });
|
||||
}
|
||||
} finally {
|
||||
fetchWithSsrFGuardMock.mockReset();
|
||||
@@ -1530,10 +1516,12 @@ describe("createOllamaStreamFn streaming events", () => {
|
||||
const types = events.map((e) => e.type);
|
||||
expect(types).toEqual(["start", "text_start", "text_delta", "error"]);
|
||||
const errorEvent = events.at(-1);
|
||||
expect(errorEvent?.type).toBe("error");
|
||||
if (errorEvent?.type === "error") {
|
||||
expect(errorEvent.error.errorMessage).toContain("garbled visible text");
|
||||
}
|
||||
expect(errorEvent).toMatchObject({
|
||||
type: "error",
|
||||
error: expect.objectContaining({
|
||||
errorMessage: expect.stringContaining("garbled visible text"),
|
||||
}),
|
||||
});
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -1581,7 +1569,7 @@ describe("createOllamaStreamFn streaming events", () => {
|
||||
expect(types).toEqual(["start", "text_start", "text_delta", "text_end", "done"]);
|
||||
|
||||
const delta = events.find((e) => e.type === "text_delta");
|
||||
expect(delta?.delta).toBe("one shot");
|
||||
expect(delta).toMatchObject({ delta: "one shot" });
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -1621,7 +1609,9 @@ describe("createOllamaStreamFn", () => {
|
||||
if (!requestBody.options) {
|
||||
throw new Error("Expected Ollama request options");
|
||||
}
|
||||
expect(requestBody.options?.num_ctx).toBeUndefined();
|
||||
// Catalog `contextWindow` flows through as `num_ctx` so the request
|
||||
// does not silently truncate to Ollama's small Modelfile default.
|
||||
expect(requestBody.options?.num_ctx).toBe(131072);
|
||||
expect(requestBody.options.num_predict).toBe(123);
|
||||
},
|
||||
);
|
||||
@@ -1704,7 +1694,7 @@ describe("createOllamaStreamFn", () => {
|
||||
);
|
||||
});
|
||||
|
||||
it("does not fall back to catalog contextWindow as native Ollama num_ctx", async () => {
|
||||
it("falls back to catalog contextWindow as num_ctx when params.num_ctx is unset", async () => {
|
||||
await withMockNdjsonFetch(
|
||||
[
|
||||
'{"model":"m","created_at":"t","message":{"role":"assistant","content":"ok"},"done":false}',
|
||||
@@ -1725,12 +1715,12 @@ describe("createOllamaStreamFn", () => {
|
||||
const requestBody = JSON.parse(requestInit.body) as {
|
||||
options?: { num_ctx?: number };
|
||||
};
|
||||
expect(requestBody.options?.num_ctx).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBe(32768);
|
||||
},
|
||||
);
|
||||
});
|
||||
|
||||
it("does not fall back to catalog maxTokens as native Ollama num_ctx", async () => {
|
||||
it("falls back to catalog maxTokens as num_ctx when contextWindow is absent", async () => {
|
||||
await withMockNdjsonFetch(
|
||||
[
|
||||
'{"model":"m","created_at":"t","message":{"role":"assistant","content":"ok"},"done":false}',
|
||||
@@ -1754,7 +1744,7 @@ describe("createOllamaStreamFn", () => {
|
||||
const requestBody = JSON.parse(requestInit.body) as {
|
||||
options?: { num_ctx?: number };
|
||||
};
|
||||
expect(requestBody.options?.num_ctx).toBeUndefined();
|
||||
expect(requestBody.options?.num_ctx).toBe(65536);
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -1802,9 +1792,10 @@ describe("createOllamaStreamFn", () => {
|
||||
|
||||
const request = getGuardedFetchCall(fetchMock);
|
||||
expect(request.url).toBe("http://127.0.0.1:11434/api/chat");
|
||||
const policy = requireRecord(request.policy, "ssrf policy");
|
||||
expect(policy.hostnameAllowlist).toEqual(["127.0.0.1"]);
|
||||
expect(policy.allowPrivateNetwork).toBe(true);
|
||||
expect(request.policy).toMatchObject({
|
||||
hostnameAllowlist: ["127.0.0.1"],
|
||||
allowPrivateNetwork: true,
|
||||
});
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -1834,11 +1825,12 @@ describe("createOllamaStreamFn", () => {
|
||||
expect(events.at(-1)?.type).toBe("done");
|
||||
|
||||
const requestInit = getGuardedFetchCall(fetchMock).init ?? {};
|
||||
const headers = requireHeaders(requestInit.headers);
|
||||
expect(headers["Content-Type"]).toBe("application/json");
|
||||
expect(headers["X-OLLAMA-KEY"]).toBe("provider-secret");
|
||||
expect(headers["X-Trace"]).toBe("request");
|
||||
expect(headers["X-Request-Only"]).toBe("1");
|
||||
expect(requestInit.headers).toMatchObject({
|
||||
"Content-Type": "application/json",
|
||||
"X-OLLAMA-KEY": "provider-secret",
|
||||
"X-Trace": "request",
|
||||
"X-Request-Only": "1",
|
||||
});
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -1865,7 +1857,9 @@ describe("createOllamaStreamFn", () => {
|
||||
|
||||
await collectStreamEvents(stream);
|
||||
const requestInit = getGuardedFetchCall(fetchMock).init ?? {};
|
||||
expect(requireHeaders(requestInit.headers).Authorization).toBe("Bearer proxy-token");
|
||||
expect(requestInit.headers).toMatchObject({
|
||||
Authorization: "Bearer proxy-token",
|
||||
});
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -1899,7 +1893,9 @@ describe("createOllamaStreamFn", () => {
|
||||
|
||||
await collectStreamEvents(stream);
|
||||
const requestInit = getGuardedFetchCall(fetchMock).init ?? {};
|
||||
expect(requireHeaders(requestInit.headers).Authorization).toBe("Bearer real-token");
|
||||
expect(requestInit.headers).toMatchObject({
|
||||
Authorization: "Bearer real-token",
|
||||
});
|
||||
},
|
||||
);
|
||||
});
|
||||
@@ -2041,7 +2037,9 @@ describe("createConfiguredOllamaStreamFn", () => {
|
||||
const request = getGuardedFetchCall(fetchMock);
|
||||
expect(request.url).toBe("http://provider-host:11434/api/chat");
|
||||
const requestInit = request.init ?? {};
|
||||
expect(requireHeaders(requestInit.headers).Authorization).toBe("Bearer proxy-token");
|
||||
expect(requestInit.headers).toMatchObject({
|
||||
Authorization: "Bearer proxy-token",
|
||||
});
|
||||
},
|
||||
);
|
||||
});
|
||||
|
||||
@@ -293,16 +293,29 @@ function resolveOllamaNumCtx(model: ProviderRuntimeModel): number {
|
||||
/**
|
||||
* Resolves num_ctx for native /api/chat requests:
|
||||
* 1. explicit `params.num_ctx` set on the model wins,
|
||||
* 2. otherwise return undefined so Ollama's model, OLLAMA_CONTEXT_LENGTH,
|
||||
* VRAM, or Modelfile policy decides.
|
||||
* 2. otherwise the catalog `contextWindow` / `maxTokens` is forwarded so
|
||||
* OpenClaw's known model windows survive the trip and `/api/chat` does
|
||||
* not silently truncate to Ollama's small Modelfile default (typically
|
||||
* 2048 tokens) — which is too small for a system prompt plus tool
|
||||
* definitions and produces "model picks wrong tools / says nonsense"
|
||||
* symptoms on agent turns,
|
||||
* 3. when neither is known, return undefined so the Modelfile decides.
|
||||
*
|
||||
* This intentionally differs from `resolveOllamaNumCtx` by not falling back
|
||||
* to `DEFAULT_CONTEXT_TOKENS`: that constant is a sane wrapper-side guess for
|
||||
* the OpenAI-compat path, but native `/api/chat` should not force the full
|
||||
* advertised catalog context for local models unless the operator opted in.
|
||||
* the OpenAI-compat path, but on the native path we prefer to leave num_ctx
|
||||
* absent rather than guess a window for an unknown model.
|
||||
*/
|
||||
function resolveOllamaNativeNumCtx(model: ProviderRuntimeModel): number | undefined {
|
||||
return resolveOllamaConfiguredNumCtx(model);
|
||||
const configured = resolveOllamaConfiguredNumCtx(model);
|
||||
if (configured !== undefined) {
|
||||
return configured;
|
||||
}
|
||||
const catalog = model.contextWindow ?? model.maxTokens;
|
||||
if (typeof catalog === "number" && Number.isFinite(catalog) && catalog > 0) {
|
||||
return Math.floor(catalog);
|
||||
}
|
||||
return undefined;
|
||||
}
|
||||
|
||||
function resolveOllamaModelOptions(model: ProviderRuntimeModel): Record<string, unknown> {
|
||||
|
||||
@@ -60,38 +60,6 @@ async function expectResponsesJson<T>(server: { baseUrl: string }, body: unknown
|
||||
return (await response.json()) as T;
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): unknown[] {
|
||||
if (!Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value;
|
||||
}
|
||||
|
||||
function outputItem(payload: unknown, index = 0) {
|
||||
const output = requireArray(requireRecord(payload, "response payload").output, "response output");
|
||||
return requireRecord(output[index], `response output ${index}`);
|
||||
}
|
||||
|
||||
function outputContentItem(payload: unknown, outputIndex = 0, contentIndex = 0) {
|
||||
const content = requireArray(outputItem(payload, outputIndex).content, "response output content");
|
||||
return requireRecord(content[contentIndex], `response content ${contentIndex}`);
|
||||
}
|
||||
|
||||
function outputText(payload: unknown, outputIndex = 0, contentIndex = 0) {
|
||||
const text = outputContentItem(payload, outputIndex, contentIndex).text;
|
||||
if (typeof text !== "string") {
|
||||
throw new Error("Expected response output text");
|
||||
}
|
||||
return text;
|
||||
}
|
||||
|
||||
function makeUserInput(text: string) {
|
||||
return {
|
||||
role: "user" as const,
|
||||
@@ -172,9 +140,18 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(preActionResponse.status).toBe(200);
|
||||
const preActionPayload = await preActionResponse.json();
|
||||
expect(outputItem(preActionPayload).type).toBe("message");
|
||||
expect(outputText(preActionPayload)).toContain("Protocol note: acknowledged.");
|
||||
expect(await preActionResponse.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "message",
|
||||
content: [
|
||||
{
|
||||
text: expect.stringContaining("Protocol note: acknowledged."),
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
|
||||
const approvalResponse = await fetch(`${server.baseUrl}/v1/responses`, {
|
||||
method: "POST",
|
||||
@@ -201,13 +178,13 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debugResponse = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debugResponse.status).toBe(200);
|
||||
const debugPayload = requireRecord(await debugResponse.json(), "debug request");
|
||||
expect(debugPayload.model).toBe("gpt-5.5");
|
||||
expect(debugPayload.prompt).toBe(
|
||||
"ok do it. read `QA_KICKOFF_TASK.md` now and reply with the QA mission in one short sentence.",
|
||||
);
|
||||
expect(String(debugPayload.allInputText)).toContain("ok do it.");
|
||||
expect(debugPayload.plannedToolName).toBe("read");
|
||||
expect(await debugResponse.json()).toMatchObject({
|
||||
model: "gpt-5.5",
|
||||
prompt:
|
||||
"ok do it. read `QA_KICKOFF_TASK.md` now and reply with the QA mission in one short sentence.",
|
||||
allInputText: expect.stringContaining("ok do it."),
|
||||
plannedToolName: "read",
|
||||
});
|
||||
});
|
||||
|
||||
it("emits deterministic text deltas for generic streaming QA prompts", async () => {
|
||||
@@ -507,14 +484,11 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debugResponse = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debugResponse.status).toBe(200);
|
||||
const debugPayload = requireRecord(await debugResponse.json(), "debug request");
|
||||
expect(debugPayload.prompt).toBe(
|
||||
'Please inspect "message_id" metadata first, then read `./QA_KICKOFF_TASK.md`.',
|
||||
);
|
||||
expect(debugPayload.allInputText).toBe(
|
||||
'Please inspect "message_id" metadata first, then read `./QA_KICKOFF_TASK.md`.',
|
||||
);
|
||||
expect(debugPayload.plannedToolName).toBe("read");
|
||||
expect(await debugResponse.json()).toMatchObject({
|
||||
prompt: 'Please inspect "message_id" metadata first, then read `./QA_KICKOFF_TASK.md`.',
|
||||
allInputText: 'Please inspect "message_id" metadata first, then read `./QA_KICKOFF_TASK.md`.',
|
||||
plannedToolName: "read",
|
||||
});
|
||||
});
|
||||
|
||||
it("drives the Lobster Invaders write flow and memory recall responses", async () => {
|
||||
@@ -591,9 +565,10 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const requests = await fetch(`${server.baseUrl}/debug/requests`);
|
||||
expect(requests.status).toBe(200);
|
||||
const requestLog = requireArray(await requests.json(), "debug requests");
|
||||
expect(requireRecord(requestLog[0], "debug request 0").model).toBe("gpt-5.5");
|
||||
expect(requireRecord(requestLog[1], "debug request 1").model).toBe("gpt-5.5-alt");
|
||||
expect((await requests.json()) as Array<{ model?: string }>).toMatchObject([
|
||||
{ model: "gpt-5.5" },
|
||||
{ model: "gpt-5.5-alt" },
|
||||
]);
|
||||
});
|
||||
|
||||
it("keeps remember prompts prose-only even when they mention repo cleanup", async () => {
|
||||
@@ -1115,13 +1090,15 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debugResponse = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debugResponse.status).toBe(200);
|
||||
const debugPayload = requireRecord(await debugResponse.json(), "debug request");
|
||||
expect(debugPayload.plannedToolName).toBe("sessions_spawn");
|
||||
const plannedToolArgs = requireRecord(debugPayload.plannedToolArgs, "planned tool args");
|
||||
expect(plannedToolArgs.task).toBe("Report the visible code");
|
||||
expect(plannedToolArgs.label).toBe("qa-fork-context");
|
||||
expect(plannedToolArgs.context).toBe("fork");
|
||||
expect(plannedToolArgs.mode).toBe("run");
|
||||
expect(await debugResponse.json()).toMatchObject({
|
||||
plannedToolName: "sessions_spawn",
|
||||
plannedToolArgs: {
|
||||
task: "Report the visible code",
|
||||
label: "qa-fork-context",
|
||||
context: "fork",
|
||||
mode: "run",
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("drives yielded-parent subagent fallback QA through sessions_spawn and sessions_yield", async () => {
|
||||
@@ -1135,15 +1112,16 @@ describe("qa mock openai server", () => {
|
||||
input: [makeUserInput(prompt)],
|
||||
});
|
||||
|
||||
const spawnDebug = requireRecord(
|
||||
await (await fetch(`${server.baseUrl}/debug/last-request`)).json(),
|
||||
"spawn debug request",
|
||||
);
|
||||
expect(spawnDebug.plannedToolName).toBe("sessions_spawn");
|
||||
const spawnArgs = requireRecord(spawnDebug.plannedToolArgs, "spawn planned tool args");
|
||||
expect(spawnArgs.label).toBe("qa-direct-fallback-worker");
|
||||
expect(spawnArgs.thread).toBe(false);
|
||||
expect(spawnArgs.mode).toBe("run");
|
||||
await expect(
|
||||
(await fetch(`${server.baseUrl}/debug/last-request`)).json(),
|
||||
).resolves.toMatchObject({
|
||||
plannedToolName: "sessions_spawn",
|
||||
plannedToolArgs: {
|
||||
label: "qa-direct-fallback-worker",
|
||||
thread: false,
|
||||
mode: "run",
|
||||
},
|
||||
});
|
||||
|
||||
const body = await expectResponsesText(server, {
|
||||
stream: true,
|
||||
@@ -1164,11 +1142,11 @@ describe("qa mock openai server", () => {
|
||||
|
||||
expect(body).toContain('"name":"sessions_yield"');
|
||||
expect(body).toContain("QA-SUBAGENT-DIRECT-FALLBACK-OK");
|
||||
const yieldDebug = requireRecord(
|
||||
await (await fetch(`${server.baseUrl}/debug/last-request`)).json(),
|
||||
"yield debug request",
|
||||
);
|
||||
expect(yieldDebug.plannedToolName).toBe("sessions_yield");
|
||||
await expect(
|
||||
(await fetch(`${server.baseUrl}/debug/last-request`)).json(),
|
||||
).resolves.toMatchObject({
|
||||
plannedToolName: "sessions_yield",
|
||||
});
|
||||
});
|
||||
|
||||
it("returns no visible announce output for the direct fallback QA marker", async () => {
|
||||
@@ -1269,13 +1247,18 @@ describe("qa mock openai server", () => {
|
||||
const server = await startMockServer();
|
||||
const childToken = "QA_SUBAGENT_CHILD_DIRECT";
|
||||
|
||||
const childPayload = await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
input: [makeUserInput(threadSubagentTask(childToken))],
|
||||
await expect(
|
||||
expectResponsesJson<{ output?: Array<{ content?: Array<{ text?: string }> }> }>(server, {
|
||||
stream: false,
|
||||
input: [makeUserInput(threadSubagentTask(childToken))],
|
||||
}),
|
||||
).resolves.toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: childToken }],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(outputText(childPayload)).toBe(childToken);
|
||||
});
|
||||
|
||||
it("plans memory tools and serves mock image generations", async () => {
|
||||
@@ -1323,18 +1306,20 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(image.status).toBe(200);
|
||||
const imagePayload = requireRecord(await image.json(), "image response");
|
||||
const imageData = requireArray(imagePayload.data, "image data");
|
||||
expect(typeof requireRecord(imageData[0], "image data 0").b64_json).toBe("string");
|
||||
expect(await image.json()).toMatchObject({
|
||||
data: [{ b64_json: expect.any(String) }],
|
||||
});
|
||||
|
||||
const imageRequests = await fetch(`${server.baseUrl}/debug/image-generations`);
|
||||
expect(imageRequests.status).toBe(200);
|
||||
const imageRequestLog = requireArray(await imageRequests.json(), "image generation requests");
|
||||
const imageRequest = requireRecord(imageRequestLog[0], "image generation request 0");
|
||||
expect(imageRequest.model).toBe("gpt-image-1");
|
||||
expect(imageRequest.prompt).toBe("Draw a QA lighthouse");
|
||||
expect(imageRequest.n).toBe(1);
|
||||
expect(imageRequest.size).toBe("1024x1024");
|
||||
expect(await imageRequests.json()).toMatchObject([
|
||||
{
|
||||
model: "gpt-image-1",
|
||||
prompt: "Draw a QA lighthouse",
|
||||
n: 1,
|
||||
size: "1024x1024",
|
||||
},
|
||||
]);
|
||||
});
|
||||
|
||||
it("supports advanced QA memory and subagent recovery prompts", async () => {
|
||||
@@ -1666,9 +1651,10 @@ describe("qa mock openai server", () => {
|
||||
);
|
||||
const lastRequest = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(lastRequest.status).toBe(200);
|
||||
const lastRequestPayload = requireRecord(await lastRequest.json(), "last request");
|
||||
expect(String(lastRequestPayload.instructions)).toContain("<active_memory_plugin>");
|
||||
expect(String(lastRequestPayload.allInputText)).toContain("<active_memory_plugin>");
|
||||
expect(await lastRequest.json()).toMatchObject({
|
||||
instructions: expect.stringContaining("<active_memory_plugin>"),
|
||||
allInputText: expect.stringContaining("<active_memory_plugin>"),
|
||||
});
|
||||
|
||||
const spawn = await fetch(`${server.baseUrl}/v1/responses`, {
|
||||
method: "POST",
|
||||
@@ -1754,7 +1740,17 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(final.status).toBe(200);
|
||||
expect(outputText(await final.json())).toBe("subagent-1: ok\nsubagent-2: ok");
|
||||
expect(await final.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [
|
||||
{
|
||||
text: "subagent-1: ok\nsubagent-2: ok",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("completes subagent fanout from a continuation turn without tool output", async () => {
|
||||
@@ -1821,7 +1817,17 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(phaseOnlyFinal.status).toBe(200);
|
||||
expect(outputText(await phaseOnlyFinal.json())).toBe("subagent-1: ok\nsubagent-2: ok");
|
||||
expect(await phaseOnlyFinal.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [
|
||||
{
|
||||
text: "subagent-1: ok\nsubagent-2: ok",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("does not let fanout completion state hijack child worker replies", async () => {
|
||||
@@ -1885,7 +1891,17 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(childReply.status).toBe(200);
|
||||
expect(outputText(await childReply.json())).toBe("ALPHA-OK");
|
||||
expect(await childReply.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [
|
||||
{
|
||||
text: "ALPHA-OK",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("keeps subagent fanout state isolated per mock server instance", async () => {
|
||||
@@ -1963,7 +1979,13 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(response.status).toBe(200);
|
||||
expect(outputText(await response.json())).toBe("HEARTBEAT_OK");
|
||||
expect(await response.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "HEARTBEAT_OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("returns exact markers for visible and hot-installed skills", async () => {
|
||||
@@ -1996,7 +2018,13 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(visible.status).toBe(200);
|
||||
expect(outputText(await visible.json())).toBe("VISIBLE-SKILL-OK");
|
||||
expect(await visible.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "VISIBLE-SKILL-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
|
||||
const hot = await fetch(`${server.baseUrl}/v1/responses`, {
|
||||
method: "POST",
|
||||
@@ -2019,7 +2047,13 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(hot.status).toBe(200);
|
||||
expect(outputText(await hot.json())).toBe("HOT-INSTALL-OK");
|
||||
expect(await hot.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "HOT-INSTALL-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("uses the latest exact marker directive from conversation history", async () => {
|
||||
@@ -2062,7 +2096,13 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(response.status).toBe(200);
|
||||
expect(outputText(await response.json())).toBe("NEW_TOKEN");
|
||||
expect(await response.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "NEW_TOKEN" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("lets the latest exact marker prompt beat stale Telegram session_status history", async () => {
|
||||
@@ -2105,7 +2145,13 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(response.status).toBe(200);
|
||||
expect(outputText(await response.json())).toBe("QA-TELEGRAM-REPLY-CHAIN-OK");
|
||||
expect(await response.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "QA-TELEGRAM-REPLY-CHAIN-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("does not repeat stale Telegram session_status for later ordinary prompts", async () => {
|
||||
@@ -2192,7 +2238,13 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(response.status).toBe(200);
|
||||
expect(outputText(await response.json())).toBe("QA_CANARY_TEST");
|
||||
expect(await response.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "QA_CANARY_TEST" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("uses image generation directives from request context when the latest user text is generic", async () => {
|
||||
@@ -2221,10 +2273,15 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(toolPlan.status).toBe(200);
|
||||
const toolPlanOutput = outputItem(await toolPlan.json());
|
||||
expect(toolPlanOutput.type).toBe("function_call");
|
||||
expect(toolPlanOutput.name).toBe("image_generate");
|
||||
expect(String(toolPlanOutput.arguments)).toContain("qa-lighthouse.png");
|
||||
expect(await toolPlan.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "function_call",
|
||||
name: "image_generate",
|
||||
arguments: expect.stringContaining("qa-lighthouse.png"),
|
||||
},
|
||||
],
|
||||
});
|
||||
|
||||
const toolResult = await fetch(`${server.baseUrl}/v1/responses`, {
|
||||
method: "POST",
|
||||
@@ -2255,7 +2312,13 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(toolResult.status).toBe(200);
|
||||
expect(outputText(await toolResult.json())).toContain("MEDIA:/tmp/qa-lighthouse.png");
|
||||
expect(await toolResult.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: expect.stringContaining("MEDIA:/tmp/qa-lighthouse.png") }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("records image inputs and describes attached images", async () => {
|
||||
@@ -2301,8 +2364,11 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debug = await fetch(`${server.baseUrl}/debug/requests`);
|
||||
expect(debug.status).toBe(200);
|
||||
const requestLog = requireArray(await debug.json(), "debug requests");
|
||||
expect(requireRecord(requestLog[0], "debug request 0").imageInputCount).toBe(1);
|
||||
expect(await debug.json()).toMatchObject([
|
||||
expect.objectContaining({
|
||||
imageInputCount: 1,
|
||||
}),
|
||||
]);
|
||||
});
|
||||
|
||||
it("recognizes OpenAI-compatible image_url parts as image inputs", async () => {
|
||||
@@ -2346,7 +2412,9 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debug = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debug.status).toBe(200);
|
||||
expect(requireRecord(await debug.json(), "debug request").imageInputCount).toBe(1);
|
||||
expect(await debug.json()).toMatchObject({
|
||||
imageInputCount: 1,
|
||||
});
|
||||
});
|
||||
|
||||
it("handles deeply nested image input shapes without recursive traversal failure", async () => {
|
||||
@@ -2388,7 +2456,9 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debug = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debug.status).toBe(200);
|
||||
expect(requireRecord(await debug.json(), "debug request").imageInputCount).toBe(1);
|
||||
expect(await debug.json()).toMatchObject({
|
||||
imageInputCount: 1,
|
||||
});
|
||||
});
|
||||
|
||||
it("describes reattached generated images in the roundtrip flow", async () => {
|
||||
@@ -2512,7 +2582,17 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
|
||||
expect(response.status).toBe(200);
|
||||
expect(outputText(await response.json())).toContain("model switch handoff confirmed");
|
||||
expect(await response.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [
|
||||
{
|
||||
text: expect.stringContaining("model switch handoff confirmed"),
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("returns NO_REPLY for unmentioned group chatter", async () => {
|
||||
@@ -2545,7 +2625,13 @@ describe("qa mock openai server", () => {
|
||||
}),
|
||||
});
|
||||
expect(response.status).toBe(200);
|
||||
expect(outputText(await response.json())).toBe("NO_REPLY");
|
||||
expect(await response.json()).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
content: [{ text: "NO_REPLY" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
});
|
||||
|
||||
it("advertises Anthropic claude-opus-4-6 baseline model on /v1/models", async () => {
|
||||
@@ -2613,9 +2699,10 @@ describe("qa mock openai server", () => {
|
||||
|
||||
const debugResponse = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debugResponse.status).toBe(200);
|
||||
const debugPayload = requireRecord(await debugResponse.json(), "debug request");
|
||||
expect(debugPayload.model).toBe("claude-opus-4-6");
|
||||
expect(debugPayload.plannedToolName).toBe("read");
|
||||
expect(await debugResponse.json()).toMatchObject({
|
||||
model: "claude-opus-4-6",
|
||||
plannedToolName: "read",
|
||||
});
|
||||
});
|
||||
|
||||
it("preserves Anthropic /v1/messages declared tools for explicit sessions_spawn prompts", async () => {
|
||||
@@ -2656,17 +2743,20 @@ describe("qa mock openai server", () => {
|
||||
| { name: string; input: Record<string, unknown> }
|
||||
| undefined;
|
||||
expect(toolUseBlock?.name).toBe("sessions_spawn");
|
||||
expect(toolUseBlock?.input.task).toBe(threadSubagentTask("QA_SUBAGENT_CHILD_ANTHROPIC"));
|
||||
expect(toolUseBlock?.input.label).toBe("qa-thread-subagent");
|
||||
expect(toolUseBlock?.input.thread).toBe(true);
|
||||
expect(toolUseBlock?.input.mode).toBe("session");
|
||||
expect(toolUseBlock?.input.runTimeoutSeconds).toBe(30);
|
||||
expect(toolUseBlock?.input).toMatchObject({
|
||||
task: threadSubagentTask("QA_SUBAGENT_CHILD_ANTHROPIC"),
|
||||
label: "qa-thread-subagent",
|
||||
thread: true,
|
||||
mode: "session",
|
||||
runTimeoutSeconds: 30,
|
||||
});
|
||||
|
||||
const debugResponse = await fetch(`${server.baseUrl}/debug/last-request`);
|
||||
expect(debugResponse.status).toBe(200);
|
||||
const debugPayload = requireRecord(await debugResponse.json(), "debug request");
|
||||
expect(debugPayload.model).toBe("claude-opus-4-6");
|
||||
expect(debugPayload.plannedToolName).toBe("sessions_spawn");
|
||||
expect(await debugResponse.json()).toMatchObject({
|
||||
model: "claude-opus-4-6",
|
||||
plannedToolName: "sessions_spawn",
|
||||
});
|
||||
});
|
||||
|
||||
it("dispatches Anthropic /v1/messages tool_result follow-ups through the shared scenario logic", async () => {
|
||||
@@ -3094,86 +3184,108 @@ describe("qa mock openai server", () => {
|
||||
expect(toolPlan).toContain('"name":"read"');
|
||||
expect(toolPlan).toContain("QA_KICKOFF_TASK.md");
|
||||
|
||||
const reasoningPayload = await expectResponsesJson<{
|
||||
output?: Array<{ type?: string; id?: string; summary?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_REASONING_ONLY_RECOVERY_PROMPT),
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{ type?: string; id?: string; summary?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_REASONING_ONLY_RECOVERY_PROMPT),
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
},
|
||||
],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
type: "reasoning",
|
||||
id: "rs_mock_reasoning_recovery",
|
||||
summary: [{ text: expect.stringContaining("Need visible answer") }],
|
||||
},
|
||||
],
|
||||
});
|
||||
const reasoningOutput = outputItem(reasoningPayload);
|
||||
expect(reasoningOutput.type).toBe("reasoning");
|
||||
expect(reasoningOutput.id).toBe("rs_mock_reasoning_recovery");
|
||||
const reasoningSummary = requireArray(reasoningOutput.summary, "reasoning summary");
|
||||
expect(String(requireRecord(reasoningSummary[0], "reasoning summary 0").text)).toContain(
|
||||
"Need visible answer",
|
||||
);
|
||||
|
||||
const recoveredPayload = await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_REASONING_ONLY_RECOVERY_PROMPT),
|
||||
makeUserInput(QA_REASONING_ONLY_RETRY_INSTRUCTION),
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_REASONING_ONLY_RECOVERY_PROMPT),
|
||||
makeUserInput(QA_REASONING_ONLY_RETRY_INSTRUCTION),
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
},
|
||||
],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
content: [{ text: "REASONING-RECOVERED-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(outputText(recoveredPayload)).toBe("REASONING-RECOVERED-OK");
|
||||
|
||||
const requests = await fetch(`${server.baseUrl}/debug/requests`);
|
||||
expect(requests.status).toBe(200);
|
||||
const requestLog = requireArray(await requests.json(), "debug requests");
|
||||
expect(requireRecord(requestLog[0], "debug request 0").plannedToolName).toBe("read");
|
||||
expect(String(requireRecord(requestLog[1], "debug request 1").allInputText)).toContain(
|
||||
QA_REASONING_ONLY_RECOVERY_PROMPT,
|
||||
);
|
||||
expect(String(requireRecord(requestLog[2], "debug request 2").allInputText)).toContain(
|
||||
QA_REASONING_ONLY_RETRY_INSTRUCTION,
|
||||
);
|
||||
expect(await requests.json()).toMatchObject([
|
||||
{ plannedToolName: "read" },
|
||||
{ allInputText: expect.stringContaining(QA_REASONING_ONLY_RECOVERY_PROMPT) },
|
||||
{ allInputText: expect.stringContaining(QA_REASONING_ONLY_RETRY_INSTRUCTION) },
|
||||
]);
|
||||
});
|
||||
|
||||
it("scripts the GPT-5.5 thinking visibility switch prompts", async () => {
|
||||
const server = await startMockServer();
|
||||
|
||||
const offPayload = await expectResponsesJson<{
|
||||
output?: Array<{ type?: string; content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [makeUserInput(QA_THINKING_VISIBILITY_OFF_PROMPT)],
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{ type?: string; content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [makeUserInput(QA_THINKING_VISIBILITY_OFF_PROMPT)],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "message",
|
||||
content: [{ text: "THINKING-OFF-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(outputItem(offPayload).type).toBe("message");
|
||||
expect(outputText(offPayload)).toBe("THINKING-OFF-OK");
|
||||
|
||||
const maxPayload = await expectResponsesJson<{
|
||||
output?: Array<{
|
||||
type?: string;
|
||||
id?: string;
|
||||
summary?: Array<{ text?: string }>;
|
||||
content?: Array<{ text?: string }>;
|
||||
}>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [makeUserInput(QA_THINKING_VISIBILITY_MAX_PROMPT)],
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{
|
||||
type?: string;
|
||||
id?: string;
|
||||
summary?: Array<{ text?: string }>;
|
||||
content?: Array<{ text?: string }>;
|
||||
}>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [makeUserInput(QA_THINKING_VISIBILITY_MAX_PROMPT)],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "reasoning",
|
||||
id: "rs_mock_thinking_visibility_max",
|
||||
summary: [],
|
||||
},
|
||||
{
|
||||
type: "message",
|
||||
content: [{ text: "THINKING-MAX-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
const maxReasoning = outputItem(maxPayload);
|
||||
expect(maxReasoning.type).toBe("reasoning");
|
||||
expect(maxReasoning.id).toBe("rs_mock_thinking_visibility_max");
|
||||
expect(maxReasoning.summary).toEqual([]);
|
||||
expect(outputItem(maxPayload, 1).type).toBe("message");
|
||||
expect(outputText(maxPayload, 1)).toBe("THINKING-MAX-OK");
|
||||
});
|
||||
|
||||
it("keeps the reasoning-only side-effect path ready for no-auto-retry QA coverage", async () => {
|
||||
@@ -3187,22 +3299,23 @@ describe("qa mock openai server", () => {
|
||||
expect(toolPlan).toContain('"name":"write"');
|
||||
expect(toolPlan).toContain("reasoning-only-side-effect.txt");
|
||||
|
||||
const sideEffectPayload = await expectResponsesJson<{
|
||||
output?: Array<{ type?: string; id?: string }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_REASONING_ONLY_SIDE_EFFECT_PROMPT),
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "Successfully wrote 28 bytes to reasoning-only-side-effect.txt.",
|
||||
},
|
||||
],
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{ type?: string; id?: string }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_REASONING_ONLY_SIDE_EFFECT_PROMPT),
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "Successfully wrote 28 bytes to reasoning-only-side-effect.txt.",
|
||||
},
|
||||
],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [{ type: "reasoning", id: "rs_mock_reasoning_side_effect" }],
|
||||
});
|
||||
const sideEffectOutput = outputItem(sideEffectPayload);
|
||||
expect(sideEffectOutput.type).toBe("reasoning");
|
||||
expect(sideEffectOutput.id).toBe("rs_mock_reasoning_side_effect");
|
||||
|
||||
const requests = await fetch(`${server.baseUrl}/debug/requests`);
|
||||
expect(requests.status).toBe(200);
|
||||
@@ -3219,38 +3332,50 @@ describe("qa mock openai server", () => {
|
||||
});
|
||||
expect(toolPlan).toContain('"name":"read"');
|
||||
|
||||
const emptyPayload = await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ type?: string; text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_EMPTY_RESPONSE_RECOVERY_PROMPT),
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ type?: string; text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_EMPTY_RESPONSE_RECOVERY_PROMPT),
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
},
|
||||
],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
content: [{ type: "output_text", text: "" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
const emptyContent = outputContentItem(emptyPayload);
|
||||
expect(emptyContent.type).toBe("output_text");
|
||||
expect(emptyContent.text).toBe("");
|
||||
|
||||
const recoveredPayload = await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_EMPTY_RESPONSE_RECOVERY_PROMPT),
|
||||
makeUserInput(QA_EMPTY_RESPONSE_RETRY_INSTRUCTION),
|
||||
expect(
|
||||
await expectResponsesJson<{
|
||||
output?: Array<{ content?: Array<{ text?: string }> }>;
|
||||
}>(server, {
|
||||
stream: false,
|
||||
model: "gpt-5.5",
|
||||
input: [
|
||||
makeUserInput(QA_EMPTY_RESPONSE_RECOVERY_PROMPT),
|
||||
makeUserInput(QA_EMPTY_RESPONSE_RETRY_INSTRUCTION),
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
},
|
||||
],
|
||||
}),
|
||||
).toMatchObject({
|
||||
output: [
|
||||
{
|
||||
type: "function_call_output",
|
||||
output: "QA mission: Understand this OpenClaw repo from source + docs before acting.",
|
||||
content: [{ text: "EMPTY-RECOVERED-OK" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(outputText(recoveredPayload)).toBe("EMPTY-RECOVERED-OK");
|
||||
});
|
||||
|
||||
it("can keep emitting empty GPT turns when the single retry budget should exhaust", async () => {
|
||||
|
||||
@@ -172,7 +172,6 @@ const QA_SKILL_WORKSHOP_GIF_PROMPT_RE =
|
||||
/externally sourced animated GIF asset|animated GIF asset in a product UI/i;
|
||||
const QA_SKILL_WORKSHOP_REVIEW_PROMPT_RE = /Review transcript for durable skill updates/i;
|
||||
const QA_RELEASE_AUDIT_PROMPT_RE = /release readiness audit for the small project/i;
|
||||
const QA_TOOL_SEARCH_PROMPT_RE = /tool search qa check/i;
|
||||
|
||||
type MockScenarioState = {
|
||||
subagentFanoutPhase: number;
|
||||
@@ -673,11 +672,6 @@ function extractActiveMemorySummary(text: string) {
|
||||
return match?.[1] ? decodeXmlEntities(match[1]).trim() : null;
|
||||
}
|
||||
|
||||
function extractToolSearchTarget(text: string): string | null {
|
||||
const match = /\btarget=([A-Za-z0-9_.:-]+)\b/.exec(text);
|
||||
return match?.[1]?.trim() || null;
|
||||
}
|
||||
|
||||
function isActiveMemorySubagentPrompt(text: string) {
|
||||
return text.includes("You are a memory search agent.");
|
||||
}
|
||||
@@ -1422,22 +1416,6 @@ async function buildResponsesPayload(
|
||||
path: readTargetFromPrompt(toolProgressPrompt || prompt || allInputText),
|
||||
});
|
||||
};
|
||||
if (QA_TOOL_SEARCH_PROMPT_RE.test(allInputText) && !toolOutput) {
|
||||
const targetTool = extractToolSearchTarget(allInputText);
|
||||
if (targetTool && hasDeclaredTool(body, "tool_search_code")) {
|
||||
return buildToolCallEventsWithArgs("tool_search_code", {
|
||||
code: [
|
||||
`const hits = await openclaw.tools.search(${JSON.stringify(targetTool)}, { limit: 1 });`,
|
||||
"const match = hits.find((tool) => tool.name === " + JSON.stringify(targetTool) + ");",
|
||||
"if (!match) throw new Error('target tool not found');",
|
||||
"return await openclaw.tools.call(match.id, { marker: 'code-mode' });",
|
||||
].join("\n"),
|
||||
});
|
||||
}
|
||||
if (targetTool && hasDeclaredTool(body, targetTool)) {
|
||||
return buildToolCallEventsWithArgs(targetTool, { marker: "normal" });
|
||||
}
|
||||
}
|
||||
if (
|
||||
allInputText.includes(QA_SUBAGENT_DIRECT_FALLBACK_MARKER) &&
|
||||
/Internal task completion event/i.test(allInputText)
|
||||
|
||||
@@ -306,62 +306,6 @@ function createContext(overrides?: {
|
||||
};
|
||||
}
|
||||
|
||||
type UnknownMock = { mock: { calls: unknown[][] } };
|
||||
|
||||
function mockCallArg(mock: unknown, index: number, label: string, argIndex = 0): unknown {
|
||||
const calls = (mock as UnknownMock).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`Expected ${label} to be a mock`);
|
||||
}
|
||||
const call = calls[index];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} call ${index + 1}`);
|
||||
}
|
||||
return call[argIndex];
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object") {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function expectRecordFields(
|
||||
actual: Record<string, unknown>,
|
||||
expected: Record<string, unknown>,
|
||||
): void {
|
||||
for (const [key, value] of Object.entries(expected)) {
|
||||
expect(actual[key]).toEqual(value);
|
||||
}
|
||||
}
|
||||
|
||||
function slackInteractionPayload(callIndex = 0): Record<string, unknown> {
|
||||
const eventText = mockCallArg(enqueueSystemEventMock, callIndex, "enqueueSystemEvent");
|
||||
if (typeof eventText !== "string") {
|
||||
throw new Error("Expected Slack interaction event text");
|
||||
}
|
||||
return JSON.parse(eventText.replace("Slack interaction: ", "")) as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function chatUpdateCall(app: { client: { chat: { update: unknown } } }, callIndex = 0) {
|
||||
return requireRecord(
|
||||
mockCallArg(app.client.chat.update, callIndex, "chat.update"),
|
||||
"chat.update",
|
||||
);
|
||||
}
|
||||
|
||||
function inputByActionId(
|
||||
inputs: Array<Record<string, unknown>>,
|
||||
actionId: string,
|
||||
): Record<string, unknown> {
|
||||
const input = inputs.find((entry) => entry.actionId === actionId);
|
||||
if (!input) {
|
||||
throw new Error(`Expected input ${actionId}`);
|
||||
}
|
||||
return input;
|
||||
}
|
||||
|
||||
describe("registerSlackInteractionEvents", () => {
|
||||
beforeAll(async () => {
|
||||
({ registerSlackInteractionEvents } = await import("./interactions.js"));
|
||||
@@ -427,10 +371,21 @@ describe("registerSlackInteractionEvents", () => {
|
||||
|
||||
expect(ack).toHaveBeenCalled();
|
||||
expect(enqueueSystemEventMock).toHaveBeenCalledTimes(1);
|
||||
const eventText = mockCallArg(enqueueSystemEventMock, 0, "enqueueSystemEvent");
|
||||
expect(typeof eventText === "string" && eventText.startsWith("Slack interaction: ")).toBe(true);
|
||||
const payload = slackInteractionPayload();
|
||||
expectRecordFields(payload, {
|
||||
const [eventText] = enqueueSystemEventMock.mock.calls[0] as [string];
|
||||
expect(eventText.startsWith("Slack interaction: ")).toBe(true);
|
||||
const payload = JSON.parse(eventText.replace("Slack interaction: ", "")) as {
|
||||
actionId: string;
|
||||
actionType: string;
|
||||
value: string;
|
||||
userId: string;
|
||||
teamId?: string;
|
||||
triggerId?: string;
|
||||
responseUrl?: string;
|
||||
channelId: string;
|
||||
messageTs: string;
|
||||
threadTs?: string;
|
||||
};
|
||||
expect(payload).toMatchObject({
|
||||
actionId: "openclaw:verify",
|
||||
actionType: "button",
|
||||
value: "approved",
|
||||
@@ -519,7 +474,7 @@ describe("registerSlackInteractionEvents", () => {
|
||||
}) => Promise<unknown>;
|
||||
}
|
||||
| undefined;
|
||||
expectRecordFields(requireRecord(dispatchCall, "dispatch call"), {
|
||||
expect(dispatchCall).toMatchObject({
|
||||
channel: "slack",
|
||||
data: "codex:approve:thread-1",
|
||||
dedupeId: "U123:C1:100.200:123.trigger:codex:approve:thread-1",
|
||||
@@ -530,24 +485,24 @@ describe("registerSlackInteractionEvents", () => {
|
||||
namespace: "codex",
|
||||
payload: "approve:thread-1",
|
||||
});
|
||||
const registrationCtx = requireRecord(
|
||||
mockCallArg(registrationHandler, 0, "registration handler"),
|
||||
"registration handler ctx",
|
||||
expect(registrationHandler).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
accountId: ctx.accountId,
|
||||
conversationId: "C1",
|
||||
interactionId: "U123:C1:100.200:123.trigger:codex:approve:thread-1",
|
||||
threadId: "100.100",
|
||||
auth: expect.objectContaining({
|
||||
isAuthorizedSender: true,
|
||||
}),
|
||||
interaction: expect.objectContaining({
|
||||
actionId: "codex",
|
||||
value: "approve:thread-1",
|
||||
data: "codex:approve:thread-1",
|
||||
namespace: "codex",
|
||||
payload: "approve:thread-1",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectRecordFields(registrationCtx, {
|
||||
accountId: ctx.accountId,
|
||||
conversationId: "C1",
|
||||
interactionId: "U123:C1:100.200:123.trigger:codex:approve:thread-1",
|
||||
threadId: "100.100",
|
||||
});
|
||||
expect(requireRecord(registrationCtx.auth, "registration auth").isAuthorizedSender).toBe(true);
|
||||
expectRecordFields(requireRecord(registrationCtx.interaction, "registration interaction"), {
|
||||
actionId: "codex",
|
||||
value: "approve:thread-1",
|
||||
data: "codex:approve:thread-1",
|
||||
namespace: "codex",
|
||||
payload: "approve:thread-1",
|
||||
});
|
||||
expect(enqueueSystemEventMock).not.toHaveBeenCalled();
|
||||
expect(app.client.chat.update).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -614,11 +569,13 @@ describe("registerSlackInteractionEvents", () => {
|
||||
payload: "approve:thread-1",
|
||||
});
|
||||
|
||||
const registrationCtx = requireRecord(
|
||||
mockCallArg(registrationHandler, 0, "registration handler"),
|
||||
"registration handler ctx",
|
||||
expect(registrationHandler).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
isAuthorizedSender: false,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(requireRecord(registrationCtx.auth, "registration auth").isAuthorizedSender).toBe(false);
|
||||
});
|
||||
|
||||
it("passes true command auth to Slack plugin interactions for allowlisted senders", async () => {
|
||||
@@ -683,11 +640,13 @@ describe("registerSlackInteractionEvents", () => {
|
||||
payload: "approve:thread-1",
|
||||
});
|
||||
|
||||
const registrationCtx = requireRecord(
|
||||
mockCallArg(registrationHandler, 0, "registration handler"),
|
||||
"registration handler ctx",
|
||||
expect(registrationHandler).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
isAuthorizedSender: true,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(requireRecord(registrationCtx.auth, "registration auth").isAuthorizedSender).toBe(true);
|
||||
});
|
||||
|
||||
it("treats Slack reply buttons as plain interaction events instead of plugin dispatch", async () => {
|
||||
@@ -726,14 +685,9 @@ describe("registerSlackInteractionEvents", () => {
|
||||
|
||||
expect(ack).toHaveBeenCalled();
|
||||
expect(dispatchPluginInteractiveHandlerMock).not.toHaveBeenCalled();
|
||||
const eventText = mockCallArg(enqueueSystemEventMock, 0, "enqueueSystemEvent");
|
||||
expect(eventText).toContain('"actionId":"openclaw:reply_button"');
|
||||
expectRecordFields(
|
||||
requireRecord(
|
||||
mockCallArg(enqueueSystemEventMock, 0, "enqueueSystemEvent", 1),
|
||||
"event options",
|
||||
),
|
||||
{
|
||||
expect(enqueueSystemEventMock).toHaveBeenCalledWith(
|
||||
expect.stringContaining('"actionId":"openclaw:reply_button"'),
|
||||
expect.objectContaining({
|
||||
contextKey: "slack:interaction:C1:100.200:openclaw:reply_button",
|
||||
deliveryContext: {
|
||||
accountId: "default",
|
||||
@@ -743,7 +697,7 @@ describe("registerSlackInteractionEvents", () => {
|
||||
},
|
||||
sessionKey: "agent:ops:slack:channel:C1",
|
||||
trusted: false,
|
||||
},
|
||||
}),
|
||||
);
|
||||
expect(resolveSessionKey).toHaveBeenCalledWith({
|
||||
channelId: "C1",
|
||||
@@ -897,12 +851,14 @@ describe("registerSlackInteractionEvents", () => {
|
||||
senderId: "U123",
|
||||
});
|
||||
expect(dispatchPluginInteractiveHandlerMock).not.toHaveBeenCalled();
|
||||
expectRecordFields(chatUpdateCall(app), {
|
||||
channel: "C1",
|
||||
ts: "100.200",
|
||||
text: "Approve this bind?",
|
||||
blocks: [],
|
||||
});
|
||||
expect(app.client.chat.update).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "C1",
|
||||
ts: "100.200",
|
||||
text: "Approve this bind?",
|
||||
blocks: [],
|
||||
}),
|
||||
);
|
||||
expect(respond).toHaveBeenCalledWith({
|
||||
text: "Binding updated.",
|
||||
response_type: "ephemeral",
|
||||
@@ -958,12 +914,14 @@ describe("registerSlackInteractionEvents", () => {
|
||||
expect(resolvePluginConversationBindingApprovalMock).not.toHaveBeenCalled();
|
||||
expect(dispatchPluginInteractiveHandlerMock).not.toHaveBeenCalled();
|
||||
expect(enqueueSystemEventMock).not.toHaveBeenCalled();
|
||||
expectRecordFields(chatUpdateCall(app), {
|
||||
channel: "C1",
|
||||
ts: "100.200",
|
||||
text: "Exec approval required",
|
||||
blocks: [],
|
||||
});
|
||||
expect(app.client.chat.update).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "C1",
|
||||
ts: "100.200",
|
||||
text: "Exec approval required",
|
||||
blocks: [],
|
||||
}),
|
||||
);
|
||||
expect(respond).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
@@ -1186,16 +1144,18 @@ describe("registerSlackInteractionEvents", () => {
|
||||
expect(payload.selectedValues).toEqual(["canary"]);
|
||||
expect(payload.selectedLabels).toEqual(["Canary"]);
|
||||
expect(app.client.chat.update).toHaveBeenCalledTimes(1);
|
||||
expectRecordFields(chatUpdateCall(app), {
|
||||
channel: "C1",
|
||||
ts: "111.222",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [{ type: "mrkdwn", text: ":white_check_mark: *Canary* selected by <@U555>" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(app.client.chat.update).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "C1",
|
||||
ts: "111.222",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [{ type: "mrkdwn", text: ":white_check_mark: *Canary* selected by <@U555>" }],
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("blocks block actions from users outside configured channel users allowlist", async () => {
|
||||
@@ -1483,21 +1443,23 @@ describe("registerSlackInteractionEvents", () => {
|
||||
});
|
||||
|
||||
expect(ack).toHaveBeenCalled();
|
||||
expectRecordFields(chatUpdateCall(app), {
|
||||
channel: "C1",
|
||||
ts: "111.223",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{
|
||||
type: "mrkdwn",
|
||||
text: ":white_check_mark: *Canary\\_\\*\\`\\~<&>* selected by <@U556>",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(app.client.chat.update).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "C1",
|
||||
ts: "111.223",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{
|
||||
type: "mrkdwn",
|
||||
text: ":white_check_mark: *Canary\\_\\*\\`\\~<&>* selected by <@U556>",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("falls back to container channel and message timestamps", async () => {
|
||||
@@ -1538,7 +1500,7 @@ describe("registerSlackInteractionEvents", () => {
|
||||
threadTs?: string;
|
||||
teamId?: string;
|
||||
};
|
||||
expectRecordFields(payload as unknown as Record<string, unknown>, {
|
||||
expect(payload).toMatchObject({
|
||||
channelId: "C222",
|
||||
messageTs: "222.333",
|
||||
threadTs: "222.111",
|
||||
@@ -1586,21 +1548,23 @@ describe("registerSlackInteractionEvents", () => {
|
||||
|
||||
expect(ack).toHaveBeenCalled();
|
||||
expect(app.client.chat.update).toHaveBeenCalledTimes(1);
|
||||
expectRecordFields(chatUpdateCall(app), {
|
||||
channel: "C2",
|
||||
ts: "333.444",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{
|
||||
type: "mrkdwn",
|
||||
text: ":white_check_mark: *Alpha, Beta, Gamma +1* selected by <@U222>",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(app.client.chat.update).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
channel: "C2",
|
||||
ts: "333.444",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{
|
||||
type: "mrkdwn",
|
||||
text: ":white_check_mark: *Alpha, Beta, Gamma +1* selected by <@U222>",
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("renders date/time/datetime picker selections in confirmation rows", async () => {
|
||||
@@ -1695,42 +1659,56 @@ describe("registerSlackInteractionEvents", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const firstUpdate = chatUpdateCall(app, 0);
|
||||
const firstBlocks = firstUpdate.blocks as unknown[];
|
||||
expectRecordFields(firstUpdate, { channel: "C3", ts: "555.666" });
|
||||
expect(firstBlocks).toHaveLength(3);
|
||||
expect(firstBlocks[0]).toEqual({
|
||||
type: "context",
|
||||
elements: [{ type: "mrkdwn", text: ":white_check_mark: *2026-02-16* selected by <@U333>" }],
|
||||
});
|
||||
|
||||
expectRecordFields(chatUpdateCall(app, 1), {
|
||||
channel: "C3",
|
||||
ts: "555.667",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [{ type: "mrkdwn", text: ":white_check_mark: *14:30* selected by <@U333>" }],
|
||||
},
|
||||
],
|
||||
});
|
||||
expectRecordFields(chatUpdateCall(app, 2), {
|
||||
channel: "C3",
|
||||
ts: "555.668",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{
|
||||
type: "mrkdwn",
|
||||
text: `:white_check_mark: *${new Date(
|
||||
selectedDateTimeEpoch * 1000,
|
||||
).toISOString()}* selected by <@U333>`,
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(app.client.chat.update).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
expect.objectContaining({
|
||||
channel: "C3",
|
||||
ts: "555.666",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{ type: "mrkdwn", text: ":white_check_mark: *2026-02-16* selected by <@U333>" },
|
||||
],
|
||||
},
|
||||
expect.anything(),
|
||||
expect.anything(),
|
||||
],
|
||||
}),
|
||||
);
|
||||
expect(app.client.chat.update).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
expect.objectContaining({
|
||||
channel: "C3",
|
||||
ts: "555.667",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [{ type: "mrkdwn", text: ":white_check_mark: *14:30* selected by <@U333>" }],
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
expect(app.client.chat.update).toHaveBeenNthCalledWith(
|
||||
3,
|
||||
expect.objectContaining({
|
||||
channel: "C3",
|
||||
ts: "555.668",
|
||||
blocks: [
|
||||
{
|
||||
type: "context",
|
||||
elements: [
|
||||
{
|
||||
type: "mrkdwn",
|
||||
text: `:white_check_mark: *${new Date(
|
||||
selectedDateTimeEpoch * 1000,
|
||||
).toISOString()}* selected by <@U333>`,
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("captures expanded selection and temporal payload fields", async () => {
|
||||
@@ -1838,7 +1816,7 @@ describe("registerSlackInteractionEvents", () => {
|
||||
teamId?: string;
|
||||
channelId?: string;
|
||||
};
|
||||
expectRecordFields(payload as unknown as Record<string, unknown>, {
|
||||
expect(payload).toMatchObject({
|
||||
actionType: "workflow_button",
|
||||
workflowTriggerUrl: "[redacted]",
|
||||
workflowId: "Wf12345",
|
||||
@@ -1925,7 +1903,7 @@ describe("registerSlackInteractionEvents", () => {
|
||||
isStackedView?: boolean;
|
||||
inputs: Array<{ actionId: string; selectedValues?: string[]; inputValue?: string }>;
|
||||
};
|
||||
expectRecordFields(payload as unknown as Record<string, unknown>, {
|
||||
expect(payload).toMatchObject({
|
||||
interactionType: "view_submission",
|
||||
actionId: "view:openclaw:deploy_form",
|
||||
callbackId: "openclaw:deploy_form",
|
||||
@@ -1938,10 +1916,12 @@ describe("registerSlackInteractionEvents", () => {
|
||||
viewHash: "[redacted]",
|
||||
isStackedView: true,
|
||||
});
|
||||
const envInput = payload.inputs.find((input) => input.actionId === "env_select");
|
||||
const notesInput = payload.inputs.find((input) => input.actionId === "notes_input");
|
||||
expect(envInput?.selectedValues).toEqual(["prod"]);
|
||||
expect(notesInput?.inputValue).toBe("ship now");
|
||||
expect(payload.inputs).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ actionId: "env_select", selectedValues: ["prod"] }),
|
||||
expect.objectContaining({ actionId: "notes_input", inputValue: "ship now" }),
|
||||
]),
|
||||
);
|
||||
expect(trackEvent).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
@@ -2165,62 +2145,75 @@ describe("registerSlackInteractionEvents", () => {
|
||||
richTextPreview?: string;
|
||||
}>;
|
||||
};
|
||||
const inputs = payload.inputs as Array<Record<string, unknown>>;
|
||||
expectRecordFields(inputByActionId(inputs, "env_select"), {
|
||||
selectedValues: ["prod"],
|
||||
selectedLabels: ["Production"],
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "assignee_select"), {
|
||||
selectedValues: ["U900"],
|
||||
selectedUsers: ["U900"],
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "channel_select"), {
|
||||
selectedValues: ["C900"],
|
||||
selectedChannels: ["C900"],
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "convo_select"), {
|
||||
selectedValues: ["G900"],
|
||||
selectedConversations: ["G900"],
|
||||
});
|
||||
expect(inputByActionId(inputs, "date_select").selectedDate).toBe("2026-02-16");
|
||||
expect(inputByActionId(inputs, "time_select").selectedTime).toBe("12:45");
|
||||
expect(inputByActionId(inputs, "datetime_select").selectedDateTime).toBe(1_771_632_300);
|
||||
expectRecordFields(inputByActionId(inputs, "radio_select"), {
|
||||
selectedValues: ["blue"],
|
||||
selectedLabels: ["Blue"],
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "checks_select"), {
|
||||
selectedValues: ["a", "b"],
|
||||
selectedLabels: ["A", "B"],
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "number_input"), {
|
||||
inputKind: "number",
|
||||
inputNumber: 42.5,
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "email_input"), {
|
||||
inputKind: "email",
|
||||
inputEmail: "team@openclaw.ai",
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "url_input"), {
|
||||
inputKind: "url",
|
||||
inputUrl: "https://docs.openclaw.ai/",
|
||||
});
|
||||
expectRecordFields(inputByActionId(inputs, "richtext_input"), {
|
||||
inputKind: "rich_text",
|
||||
richTextPreview: "Ship this now with canary metrics",
|
||||
richTextValue: {
|
||||
type: "rich_text",
|
||||
elements: [
|
||||
{
|
||||
type: "rich_text_section",
|
||||
expect(payload.inputs).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
actionId: "env_select",
|
||||
selectedValues: ["prod"],
|
||||
selectedLabels: ["Production"],
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "assignee_select",
|
||||
selectedValues: ["U900"],
|
||||
selectedUsers: ["U900"],
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "channel_select",
|
||||
selectedValues: ["C900"],
|
||||
selectedChannels: ["C900"],
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "convo_select",
|
||||
selectedValues: ["G900"],
|
||||
selectedConversations: ["G900"],
|
||||
}),
|
||||
expect.objectContaining({ actionId: "date_select", selectedDate: "2026-02-16" }),
|
||||
expect.objectContaining({ actionId: "time_select", selectedTime: "12:45" }),
|
||||
expect.objectContaining({ actionId: "datetime_select", selectedDateTime: 1_771_632_300 }),
|
||||
expect.objectContaining({
|
||||
actionId: "radio_select",
|
||||
selectedValues: ["blue"],
|
||||
selectedLabels: ["Blue"],
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "checks_select",
|
||||
selectedValues: ["a", "b"],
|
||||
selectedLabels: ["A", "B"],
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "number_input",
|
||||
inputKind: "number",
|
||||
inputNumber: 42.5,
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "email_input",
|
||||
inputKind: "email",
|
||||
inputEmail: "team@openclaw.ai",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "url_input",
|
||||
inputKind: "url",
|
||||
inputUrl: "https://docs.openclaw.ai/",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
actionId: "richtext_input",
|
||||
inputKind: "rich_text",
|
||||
richTextPreview: "Ship this now with canary metrics",
|
||||
richTextValue: {
|
||||
type: "rich_text",
|
||||
elements: [
|
||||
{ type: "text", text: "Ship this now" },
|
||||
{ type: "text", text: "with canary metrics" },
|
||||
{
|
||||
type: "rich_text_section",
|
||||
elements: [
|
||||
{ type: "text", text: "Ship this now" },
|
||||
{ type: "text", text: "with canary metrics" },
|
||||
],
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
},
|
||||
});
|
||||
}),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
it("truncates rich text preview to keep payload summaries compact", async () => {
|
||||
@@ -2337,7 +2330,7 @@ describe("registerSlackInteractionEvents", () => {
|
||||
isStackedView?: boolean;
|
||||
inputs: Array<{ actionId: string; selectedValues?: string[] }>;
|
||||
};
|
||||
expectRecordFields(payload as unknown as Record<string, unknown>, {
|
||||
expect(payload).toMatchObject({
|
||||
interactionType: "view_closed",
|
||||
actionId: "view:openclaw:deploy_form",
|
||||
callbackId: "openclaw:deploy_form",
|
||||
@@ -2351,10 +2344,11 @@ describe("registerSlackInteractionEvents", () => {
|
||||
viewHash: "[redacted]",
|
||||
isStackedView: true,
|
||||
});
|
||||
expect(
|
||||
inputByActionId(payload.inputs as Array<Record<string, unknown>>, "env_select")
|
||||
.selectedValues,
|
||||
).toEqual(["canary"]);
|
||||
expect(payload.inputs).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ actionId: "env_select", selectedValues: ["canary"] }),
|
||||
]),
|
||||
);
|
||||
expect(trackEvent).toHaveBeenCalledTimes(1);
|
||||
expect(options.sessionKey).toBe("agent:main:slack:channel:C99");
|
||||
});
|
||||
|
||||
@@ -107,11 +107,6 @@ let resolveTelegramTransport: typeof import("./fetch.js").resolveTelegramTranspo
|
||||
type TelegramDispatcherPolicy = NonNullable<
|
||||
ReturnType<typeof resolveTelegramTransport>["dispatcherAttempts"]
|
||||
>[number]["dispatcherPolicy"];
|
||||
type DirectTelegramDispatcherPolicy = Extract<TelegramDispatcherPolicy, { mode: "direct" }>;
|
||||
type ExplicitProxyTelegramDispatcherPolicy = Extract<
|
||||
TelegramDispatcherPolicy,
|
||||
{ mode: "explicit-proxy" }
|
||||
>;
|
||||
|
||||
beforeAll(async () => {
|
||||
({ resolveTelegramApiBase, resolveTelegramFetch, resolveTelegramTransport } =
|
||||
@@ -226,9 +221,12 @@ function expectStickyAutoSelectDispatcher(
|
||||
| undefined,
|
||||
field: "connect" | "proxyTls" | "requestTls" = "connect",
|
||||
): void {
|
||||
const options = dispatcher?.options?.[field];
|
||||
expect(options?.autoSelectFamily).toBe(true);
|
||||
expect(options?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(dispatcher?.options?.[field]).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: true,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
}),
|
||||
);
|
||||
}
|
||||
|
||||
function expectHttp1OnlyDispatcher(
|
||||
@@ -240,7 +238,11 @@ function expectHttp1OnlyDispatcher(
|
||||
}
|
||||
| undefined,
|
||||
): void {
|
||||
expect(dispatcher?.options?.allowH2).toBe(false);
|
||||
expect(dispatcher?.options).toEqual(
|
||||
expect.objectContaining({
|
||||
allowH2: false,
|
||||
}),
|
||||
);
|
||||
}
|
||||
|
||||
function expectPinnedIpv4ConnectDispatcher(args: {
|
||||
@@ -249,8 +251,12 @@ function expectPinnedIpv4ConnectDispatcher(args: {
|
||||
followupCall?: number;
|
||||
}): void {
|
||||
const pinnedDispatcher = getDispatcherFromUndiciCall(args.pinnedCall);
|
||||
expect(pinnedDispatcher?.options?.connect?.family).toBe(4);
|
||||
expect(pinnedDispatcher?.options?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(pinnedDispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
family: 4,
|
||||
autoSelectFamily: false,
|
||||
}),
|
||||
);
|
||||
if (args.firstCall) {
|
||||
expect(getDispatcherFromUndiciCall(args.firstCall)).not.toBe(pinnedDispatcher);
|
||||
}
|
||||
@@ -261,9 +267,13 @@ function expectPinnedIpv4ConnectDispatcher(args: {
|
||||
|
||||
function expectPinnedFallbackIpDispatcher(callIndex: number) {
|
||||
const dispatcher = getDispatcherFromUndiciCall(callIndex);
|
||||
expect(dispatcher?.options?.connect?.family).toBe(4);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(typeof dispatcher?.options?.connect?.lookup).toBe("function");
|
||||
expect(dispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
family: 4,
|
||||
autoSelectFamily: false,
|
||||
lookup: expect.any(Function),
|
||||
}),
|
||||
);
|
||||
const callback = vi.fn();
|
||||
(
|
||||
dispatcher?.options?.connect?.lookup as
|
||||
@@ -358,9 +368,13 @@ describe("resolveTelegramFetch", () => {
|
||||
|
||||
const dispatcher = getDispatcherFromUndiciCall(1);
|
||||
expectHttp1OnlyDispatcher(dispatcher);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamily).toBe(true);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(typeof dispatcher?.options?.connect?.lookup).toBe("function");
|
||||
expect(dispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: true,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
lookup: expect.any(Function),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("emits default transport decisions at debug level", () => {
|
||||
@@ -386,15 +400,27 @@ describe("resolveTelegramFetch", () => {
|
||||
await resolved("https://api.telegram.org/botx/getMe");
|
||||
|
||||
expect(EnvHttpProxyAgentCtor).toHaveBeenCalledTimes(1);
|
||||
expect(EnvHttpProxyAgentCtor.mock.calls[0]?.[0]?.httpsProxy).toBe("http://127.0.0.1:7890");
|
||||
expect(EnvHttpProxyAgentCtor).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
httpsProxy: "http://127.0.0.1:7890",
|
||||
}),
|
||||
);
|
||||
expect(AgentCtor).not.toHaveBeenCalled();
|
||||
|
||||
const dispatcher = getDispatcherFromUndiciCall(1);
|
||||
expectHttp1OnlyDispatcher(dispatcher);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(dispatcher?.options?.proxyTls?.autoSelectFamily).toBe(false);
|
||||
expect(dispatcher?.options?.proxyTls?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(dispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: false,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
}),
|
||||
);
|
||||
expect(dispatcher?.options?.proxyTls).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: false,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("uses the OpenClaw debug proxy URL when no explicit proxy fetch is provided", async () => {
|
||||
@@ -406,11 +432,12 @@ describe("resolveTelegramFetch", () => {
|
||||
await resolved("https://api.telegram.org/botTOKEN/getMe");
|
||||
|
||||
expect(ProxyAgentCtor).toHaveBeenCalledTimes(1);
|
||||
const proxyOptions = ProxyAgentCtor.mock.calls[0]?.[0] as
|
||||
| { allowH2?: boolean; uri?: string }
|
||||
| undefined;
|
||||
expect(proxyOptions?.allowH2).toBe(false);
|
||||
expect(proxyOptions?.uri).toBe("http://127.0.0.1:7777");
|
||||
expect(ProxyAgentCtor).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
allowH2: false,
|
||||
uri: "http://127.0.0.1:7777",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("uses OPENCLAW_PROXY_URL as a Telegram explicit proxy when proxy env is absent", async () => {
|
||||
@@ -427,19 +454,23 @@ describe("resolveTelegramFetch", () => {
|
||||
await transport.fetch("https://api.telegram.org/botTOKEN/getMe");
|
||||
|
||||
expect(ProxyAgentCtor).toHaveBeenCalledTimes(1);
|
||||
const proxyOptions = ProxyAgentCtor.mock.calls[0]?.[0] as
|
||||
| { allowH2?: boolean; uri?: string; requestTls?: { autoSelectFamily?: boolean } }
|
||||
| undefined;
|
||||
expect(proxyOptions?.allowH2).toBe(false);
|
||||
expect(proxyOptions?.uri).toBe("http://127.0.0.1:7788");
|
||||
expect(proxyOptions?.requestTls?.autoSelectFamily).toBe(false);
|
||||
expect(ProxyAgentCtor).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
allowH2: false,
|
||||
uri: "http://127.0.0.1:7788",
|
||||
requestTls: expect.objectContaining({
|
||||
autoSelectFamily: false,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(EnvHttpProxyAgentCtor).not.toHaveBeenCalled();
|
||||
expect(AgentCtor).not.toHaveBeenCalled();
|
||||
const dispatcherPolicy = transport.dispatcherAttempts?.[0]?.dispatcherPolicy as
|
||||
| ExplicitProxyTelegramDispatcherPolicy
|
||||
| undefined;
|
||||
expect(dispatcherPolicy?.mode).toBe("explicit-proxy");
|
||||
expect(dispatcherPolicy?.proxyUrl).toBe("http://127.0.0.1:7788");
|
||||
expect(transport.dispatcherAttempts?.[0]?.dispatcherPolicy).toEqual(
|
||||
expect.objectContaining({
|
||||
mode: "explicit-proxy",
|
||||
proxyUrl: "http://127.0.0.1:7788",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("preserves caller-provided custom fetch when OPENCLAW_PROXY_URL is present", async () => {
|
||||
@@ -498,10 +529,18 @@ describe("resolveTelegramFetch", () => {
|
||||
|
||||
const dispatcher = getDispatcherFromUndiciCall(1);
|
||||
expectHttp1OnlyDispatcher(dispatcher);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamily).toBe(true);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(dispatcher?.options?.proxyTls?.autoSelectFamily).toBe(true);
|
||||
expect(dispatcher?.options?.proxyTls?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(dispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: true,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
}),
|
||||
);
|
||||
expect(dispatcher?.options?.proxyTls).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: true,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps resolver-scoped transport policy for OpenClaw proxy fetches", async () => {
|
||||
@@ -524,10 +563,16 @@ describe("resolveTelegramFetch", () => {
|
||||
expect(AgentCtor).not.toHaveBeenCalled();
|
||||
const dispatcher = getDispatcherFromUndiciCall(1);
|
||||
expectHttp1OnlyDispatcher(dispatcher);
|
||||
expect((dispatcher?.options as { uri?: string } | undefined)?.uri).toBe(
|
||||
"http://127.0.0.1:7890",
|
||||
expect(dispatcher?.options).toEqual(
|
||||
expect.objectContaining({
|
||||
uri: "http://127.0.0.1:7890",
|
||||
}),
|
||||
);
|
||||
expect(dispatcher?.options?.requestTls).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: false,
|
||||
}),
|
||||
);
|
||||
expect(dispatcher?.options?.requestTls?.autoSelectFamily).toBe(false);
|
||||
});
|
||||
|
||||
it("exports fallback dispatcher attempts for Telegram media downloads", async () => {
|
||||
@@ -550,28 +595,43 @@ describe("resolveTelegramFetch", () => {
|
||||
expect(transport.dispatcherAttempts).toHaveLength(3);
|
||||
|
||||
const [defaultAttempt, ipv4Attempt, pinnedAttempt] = transport.dispatcherAttempts as Array<{
|
||||
dispatcherPolicy?: DirectTelegramDispatcherPolicy;
|
||||
dispatcherPolicy?: TelegramDispatcherPolicy;
|
||||
}>;
|
||||
|
||||
const defaultPolicy = defaultAttempt.dispatcherPolicy;
|
||||
const ipv4Policy = ipv4Attempt.dispatcherPolicy;
|
||||
const pinnedPolicy = pinnedAttempt.dispatcherPolicy;
|
||||
expect(defaultPolicy?.mode).toBe("direct");
|
||||
expect(defaultPolicy?.connect?.autoSelectFamily).toBe(true);
|
||||
expect(defaultPolicy?.connect?.autoSelectFamilyAttemptTimeout).toBe(300);
|
||||
expect(typeof defaultPolicy?.connect?.lookup).toBe("function");
|
||||
expect(ipv4Policy?.mode).toBe("direct");
|
||||
expect(ipv4Policy?.connect?.family).toBe(4);
|
||||
expect(ipv4Policy?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(typeof ipv4Policy?.connect?.lookup).toBe("function");
|
||||
expect(pinnedPolicy?.mode).toBe("direct");
|
||||
expect(pinnedPolicy?.pinnedHostname).toEqual({
|
||||
hostname: "api.telegram.org",
|
||||
addresses: ["149.154.167.220"],
|
||||
});
|
||||
expect(pinnedPolicy?.connect?.family).toBe(4);
|
||||
expect(pinnedPolicy?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(typeof pinnedPolicy?.connect?.lookup).toBe("function");
|
||||
expect(defaultAttempt.dispatcherPolicy).toEqual(
|
||||
expect.objectContaining({
|
||||
mode: "direct",
|
||||
connect: expect.objectContaining({
|
||||
autoSelectFamily: true,
|
||||
autoSelectFamilyAttemptTimeout: 300,
|
||||
lookup: expect.any(Function),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(ipv4Attempt.dispatcherPolicy).toEqual(
|
||||
expect.objectContaining({
|
||||
mode: "direct",
|
||||
connect: expect.objectContaining({
|
||||
family: 4,
|
||||
autoSelectFamily: false,
|
||||
lookup: expect.any(Function),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(pinnedAttempt.dispatcherPolicy).toEqual(
|
||||
expect.objectContaining({
|
||||
mode: "direct",
|
||||
pinnedHostname: {
|
||||
hostname: "api.telegram.org",
|
||||
addresses: ["149.154.167.220"],
|
||||
},
|
||||
connect: expect.objectContaining({
|
||||
family: 4,
|
||||
autoSelectFamily: false,
|
||||
lookup: expect.any(Function),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("does not blind-retry when sticky IPv4 fallback is disallowed for explicit proxy paths", async () => {
|
||||
@@ -628,13 +688,20 @@ describe("resolveTelegramFetch", () => {
|
||||
await resolved("https://api.telegram.org/botx/sendMessage");
|
||||
|
||||
expect(EnvHttpProxyAgentCtor).toHaveBeenCalledTimes(1);
|
||||
const proxyOptions = EnvHttpProxyAgentCtor.mock.calls[0]?.[0];
|
||||
expect(proxyOptions?.allowH2).toBe(false);
|
||||
expect(proxyOptions?.httpProxy).toBe("http://127.0.0.1:7891");
|
||||
expect(proxyOptions?.httpsProxy).toBe("http://127.0.0.1:7891");
|
||||
expect(EnvHttpProxyAgentCtor).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
allowH2: false,
|
||||
httpProxy: "http://127.0.0.1:7891",
|
||||
httpsProxy: "http://127.0.0.1:7891",
|
||||
}),
|
||||
);
|
||||
expect(AgentCtor).not.toHaveBeenCalled();
|
||||
|
||||
expect(transport.dispatcherAttempts?.[0]?.dispatcherPolicy?.mode).toBe("env-proxy");
|
||||
expect(transport.dispatcherAttempts?.[0]?.dispatcherPolicy).toEqual(
|
||||
expect.objectContaining({
|
||||
mode: "env-proxy",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("arms sticky IPv4 fallback when env proxy init falls back to direct Agent", async () => {
|
||||
@@ -727,7 +794,11 @@ describe("resolveTelegramFetch", () => {
|
||||
expect(AgentCtor).toHaveBeenCalledTimes(1);
|
||||
|
||||
const dispatcher = getDispatcherFromUndiciCall(1);
|
||||
expect(dispatcher?.options?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(dispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: false,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("retries once, keeps sticky IPv4, then recovers to primary dispatcher", async () => {
|
||||
@@ -763,8 +834,12 @@ describe("resolveTelegramFetch", () => {
|
||||
expect(eighthDispatcher).toBe(firstDispatcher);
|
||||
|
||||
expectStickyAutoSelectDispatcher(firstDispatcher);
|
||||
expect(secondDispatcher?.options?.connect?.family).toBe(4);
|
||||
expect(secondDispatcher?.options?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(secondDispatcher?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
family: 4,
|
||||
autoSelectFamily: false,
|
||||
}),
|
||||
);
|
||||
expect(loggerDebug).toHaveBeenCalledWith(
|
||||
expect.stringContaining("fetch fallback: enabling sticky IPv4-only dispatcher"),
|
||||
);
|
||||
@@ -1026,8 +1101,16 @@ describe("resolveTelegramFetch", () => {
|
||||
|
||||
expect(dispatcherA).not.toBe(dispatcherB);
|
||||
|
||||
expect(dispatcherA?.options?.connect?.autoSelectFamily).toBe(false);
|
||||
expect(dispatcherB?.options?.connect?.autoSelectFamily).toBe(true);
|
||||
expect(dispatcherA?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: false,
|
||||
}),
|
||||
);
|
||||
expect(dispatcherB?.options?.connect).toEqual(
|
||||
expect.objectContaining({
|
||||
autoSelectFamily: true,
|
||||
}),
|
||||
);
|
||||
|
||||
// Core guarantee: Telegram transport no longer mutates process-global defaults.
|
||||
expect(setGlobalDispatcher).not.toHaveBeenCalled();
|
||||
@@ -1047,16 +1130,15 @@ describe("resolveTelegramFetch", () => {
|
||||
// One direct Agent for the default dispatcher plus two lazy fallbacks not yet touched.
|
||||
expect(AgentCtor).toHaveBeenCalledTimes(1);
|
||||
const defaultAgent = AgentCtor.mock.instances[0]?.options;
|
||||
expect(typeof defaultAgent).toBe("object");
|
||||
expect((defaultAgent as { allowH2?: boolean } | undefined)?.allowH2).toBe(false);
|
||||
expect(typeof (defaultAgent as { keepAliveTimeout?: unknown }).keepAliveTimeout).toBe(
|
||||
"number",
|
||||
expect(defaultAgent).toEqual(
|
||||
expect.objectContaining({
|
||||
allowH2: false,
|
||||
keepAliveTimeout: expect.any(Number),
|
||||
keepAliveMaxTimeout: expect.any(Number),
|
||||
connections: expect.any(Number),
|
||||
pipelining: expect.any(Number),
|
||||
}),
|
||||
);
|
||||
expect(typeof (defaultAgent as { keepAliveMaxTimeout?: unknown }).keepAliveMaxTimeout).toBe(
|
||||
"number",
|
||||
);
|
||||
expect(typeof (defaultAgent as { connections?: unknown }).connections).toBe("number");
|
||||
expect(typeof (defaultAgent as { pipelining?: unknown }).pipelining).toBe("number");
|
||||
const connections = (defaultAgent as { connections?: number }).connections;
|
||||
expect(connections).toBeGreaterThan(0);
|
||||
expect(connections).toBeLessThan(100);
|
||||
|
||||
@@ -21,7 +21,6 @@ type EmbeddedAgentArgs = {
|
||||
agentId?: string;
|
||||
workspaceDir?: string;
|
||||
sessionFile?: string;
|
||||
toolsAllow?: string[];
|
||||
};
|
||||
|
||||
function createAgentRuntime(payloads: Array<Record<string, unknown>>) {
|
||||
@@ -332,39 +331,4 @@ describe("generateVoiceResponse", () => {
|
||||
expect(args.workspaceDir).toBe("/tmp/openclaw/workspace/voice");
|
||||
expect(args.sessionFile).toBe("/tmp/openclaw/voice/sessions/session.jsonl");
|
||||
});
|
||||
|
||||
it("passes the routed voice agent explicit tool allowlist to the embedded run", async () => {
|
||||
const { runtime, runEmbeddedPiAgent } = createAgentRuntime([
|
||||
{ text: '{"spoken":"No tools needed."}' },
|
||||
]);
|
||||
const coreConfig = {
|
||||
agents: {
|
||||
list: [
|
||||
{
|
||||
id: "voice",
|
||||
tools: { allow: [] },
|
||||
},
|
||||
],
|
||||
},
|
||||
} as CoreConfig;
|
||||
|
||||
const result = await generateVoiceResponse({
|
||||
voiceConfig: VoiceCallConfigSchema.parse({
|
||||
agentId: "voice",
|
||||
responseModel: "ollama/qwen2.5:1.5b",
|
||||
responseTimeoutMs: 5000,
|
||||
}),
|
||||
coreConfig,
|
||||
agentRuntime: runtime,
|
||||
callId: "call-123",
|
||||
from: "+15550001111",
|
||||
transcript: [],
|
||||
userMessage: "hello there",
|
||||
});
|
||||
|
||||
expect(result.text).toBe("No tools needed.");
|
||||
const args = requireEmbeddedAgentArgs(runEmbeddedPiAgent);
|
||||
expect(args.agentId).toBe("voice");
|
||||
expect(args.toolsAllow).toStrictEqual([]);
|
||||
});
|
||||
});
|
||||
|
||||
@@ -41,34 +41,6 @@ type VoiceResponsePayload = {
|
||||
isReasoning?: boolean;
|
||||
};
|
||||
|
||||
function isRecord(value: unknown): value is Record<string, unknown> {
|
||||
return typeof value === "object" && value !== null && !Array.isArray(value);
|
||||
}
|
||||
|
||||
function readExplicitToolsAllow(value: unknown): string[] | undefined {
|
||||
if (!isRecord(value)) {
|
||||
return undefined;
|
||||
}
|
||||
|
||||
const allow = value.allow;
|
||||
if (!Array.isArray(allow)) {
|
||||
return undefined;
|
||||
}
|
||||
|
||||
return allow.filter((entry): entry is string => typeof entry === "string");
|
||||
}
|
||||
|
||||
function resolveVoiceAgentToolsAllow(config: CoreConfig, agentId: string): string[] | undefined {
|
||||
const agents = isRecord(config.agents) ? config.agents : undefined;
|
||||
const list = Array.isArray(agents?.list) ? agents.list : [];
|
||||
const agent = list.find((entry) => isRecord(entry) && entry.id === agentId);
|
||||
if (!isRecord(agent)) {
|
||||
return undefined;
|
||||
}
|
||||
|
||||
return readExplicitToolsAllow(isRecord(agent.tools) ? agent.tools : undefined);
|
||||
}
|
||||
|
||||
const VOICE_SPOKEN_OUTPUT_CONTRACT = [
|
||||
"Output format requirements:",
|
||||
'- Return only valid JSON in this exact shape: {"spoken":"..."}',
|
||||
@@ -240,7 +212,6 @@ export async function generateVoiceResponse(
|
||||
explicitSessionKey: sessionKey,
|
||||
});
|
||||
const agentId = voiceConfig.agentId ?? "main";
|
||||
const toolsAllow = resolveVoiceAgentToolsAllow(cfg, agentId);
|
||||
|
||||
// Resolve paths
|
||||
const storePath = agentRuntime.session.resolveStorePath(cfg.session?.store, { agentId });
|
||||
@@ -331,7 +302,6 @@ export async function generateVoiceResponse(
|
||||
lane: "voice",
|
||||
extraSystemPrompt,
|
||||
agentDir,
|
||||
toolsAllow,
|
||||
});
|
||||
|
||||
const text = extractSpokenTextFromPayloads((result.payloads ?? []) as VoiceResponsePayload[]);
|
||||
|
||||
@@ -129,53 +129,6 @@ function expectFirstSendMediaPayload(msg: WebInboundMsg) {
|
||||
return payload;
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function mockCallArg(mock: unknown, callIndex: number, argIndex: number, label: string) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`expected ${label} mock calls`);
|
||||
}
|
||||
const call = calls[callIndex];
|
||||
if (!call) {
|
||||
throw new Error(`expected ${label} call ${callIndex + 1}`);
|
||||
}
|
||||
return call[argIndex];
|
||||
}
|
||||
|
||||
function findLoggerContext(mock: unknown, message: string, label: string) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`expected ${label} mock calls`);
|
||||
}
|
||||
const call = calls.find((entry) => entry[1] === message);
|
||||
if (!call) {
|
||||
throw new Error(`expected ${label} message ${message}`);
|
||||
}
|
||||
return requireRecord(call[0], `${label} context`);
|
||||
}
|
||||
|
||||
function expectBuffer(value: unknown, label: string) {
|
||||
expect(Buffer.isBuffer(value), label).toBe(true);
|
||||
}
|
||||
|
||||
function expectQuotedOptions(
|
||||
options: unknown,
|
||||
expected: { id: string; fromMe: boolean; participant: string; body: string },
|
||||
) {
|
||||
const quoted = requireRecord(requireRecord(options, "reply options").quoted, "quoted message");
|
||||
const key = requireRecord(quoted.key, "quoted key");
|
||||
expect(key.id).toBe(expected.id);
|
||||
expect(key.fromMe).toBe(expected.fromMe);
|
||||
expect(key.participant).toBe(expected.participant);
|
||||
expect(quoted.message).toEqual({ conversation: expected.body });
|
||||
}
|
||||
|
||||
function mockSecondReplySuccess(msg: WebInboundMsg) {
|
||||
(msg.reply as unknown as { mockResolvedValueOnce: (v: unknown) => void }).mockResolvedValueOnce(
|
||||
acceptedSendResult("text", "reply-retry-2"),
|
||||
@@ -253,14 +206,21 @@ describe("deliverWebReply", () => {
|
||||
expect(msg.reply).toHaveBeenCalledTimes(2);
|
||||
expect(msg.reply).toHaveBeenNthCalledWith(1, "aaa", undefined);
|
||||
expect(msg.reply).toHaveBeenNthCalledWith(2, "aaa", undefined);
|
||||
expect(typeof mockCallArg(replyLogger.info, 0, 0, "replyLogger.info")).toBe("object");
|
||||
expect(mockCallArg(replyLogger.info, 0, 1, "replyLogger.info")).toBe("auto-reply sent (text)");
|
||||
expect(replyLogger.info).toHaveBeenCalledWith(expect.any(Object), "auto-reply sent (text)");
|
||||
expect(delivery.providerAccepted).toBe(true);
|
||||
expect(listMessageReceiptPlatformIds(delivery.receipt)).toEqual(["reply-sent-1"]);
|
||||
expect(delivery.receipt.primaryPlatformMessageId).toBe("reply-sent-1");
|
||||
expect(delivery.receipt.platformMessageIds).toEqual(["reply-sent-1"]);
|
||||
expect(delivery.receipt.parts[0]?.platformMessageId).toBe("reply-sent-1");
|
||||
expect(delivery.receipt.parts[0]?.kind).toBe("text");
|
||||
expect(delivery.receipt).toEqual(
|
||||
expect.objectContaining({
|
||||
primaryPlatformMessageId: "reply-sent-1",
|
||||
platformMessageIds: ["reply-sent-1"],
|
||||
}),
|
||||
);
|
||||
expect(delivery.receipt.parts).toEqual([
|
||||
expect.objectContaining({
|
||||
platformMessageId: "reply-sent-1",
|
||||
kind: "text",
|
||||
}),
|
||||
]);
|
||||
});
|
||||
|
||||
it("reports text replies that Baileys did not accept", async () => {
|
||||
@@ -277,11 +237,15 @@ describe("deliverWebReply", () => {
|
||||
});
|
||||
|
||||
expect(msg.reply).toHaveBeenCalledTimes(1);
|
||||
expect(delivery.receipt.platformMessageIds).toEqual([]);
|
||||
expect(delivery.receipt.parts).toEqual([]);
|
||||
expect(delivery.providerAccepted).toBe(false);
|
||||
expect(typeof mockCallArg(replyLogger.warn, 0, 0, "replyLogger.warn")).toBe("object");
|
||||
expect(mockCallArg(replyLogger.warn, 0, 1, "replyLogger.warn")).toBe(
|
||||
expect(delivery).toMatchObject({
|
||||
receipt: expect.objectContaining({
|
||||
platformMessageIds: [],
|
||||
parts: [],
|
||||
}),
|
||||
providerAccepted: false,
|
||||
});
|
||||
expect(replyLogger.warn).toHaveBeenCalledWith(
|
||||
expect.any(Object),
|
||||
"auto-reply text was not accepted by WhatsApp provider",
|
||||
);
|
||||
});
|
||||
@@ -374,20 +338,34 @@ describe("deliverWebReply", () => {
|
||||
});
|
||||
|
||||
expect(msg.reply).toHaveBeenCalledTimes(2);
|
||||
expect(mockCallArg(msg.reply, 0, 0, "reply")).toBe("aaa");
|
||||
expectQuotedOptions(mockCallArg(msg.reply, 0, 1, "reply"), {
|
||||
id: "reply-1",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
body: "quoted body",
|
||||
});
|
||||
expect(mockCallArg(msg.reply, 1, 0, "reply")).toBe("aaa");
|
||||
expectQuotedOptions(mockCallArg(msg.reply, 1, 1, "reply"), {
|
||||
id: "reply-1",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
body: "quoted body",
|
||||
});
|
||||
expect(msg.reply).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
"aaa",
|
||||
expect.objectContaining({
|
||||
quoted: expect.objectContaining({
|
||||
key: expect.objectContaining({
|
||||
id: "reply-1",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
}),
|
||||
message: { conversation: "quoted body" },
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(msg.reply).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
"aaa",
|
||||
expect.objectContaining({
|
||||
quoted: expect.objectContaining({
|
||||
key: expect.objectContaining({
|
||||
id: "reply-1",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
}),
|
||||
message: { conversation: "quoted body" },
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it.each(["connection closed", "operation timed out"])(
|
||||
@@ -449,18 +427,16 @@ describe("deliverWebReply", () => {
|
||||
localRoots: mediaLocalRoots,
|
||||
});
|
||||
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
image: expect.any(Buffer),
|
||||
caption: "aaa",
|
||||
mimetype: "image/jpeg",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(mediaPayload.image, "sendMedia image");
|
||||
expect(mediaPayload.caption).toBe("aaa");
|
||||
expect(mediaPayload.mimetype).toBe("image/jpeg");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
expect(msg.reply).toHaveBeenCalledWith("aaa", undefined);
|
||||
expect(
|
||||
findLoggerContext(replyLogger.info, "auto-reply sent (media)", "replyLogger.info"),
|
||||
).toBeDefined();
|
||||
expect(replyLogger.info).toHaveBeenCalledWith(expect.any(Object), "auto-reply sent (media)");
|
||||
expect(logVerbose).toHaveBeenCalled();
|
||||
});
|
||||
|
||||
@@ -502,26 +478,36 @@ describe("deliverWebReply", () => {
|
||||
skipLog: true,
|
||||
});
|
||||
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
image: expect.any(Buffer),
|
||||
caption: "caption",
|
||||
mimetype: "image/jpeg",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
quoted: expect.objectContaining({
|
||||
key: expect.objectContaining({
|
||||
id: "reply-2",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
}),
|
||||
message: { conversation: "quoted media body" },
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(msg.reply).toHaveBeenCalledWith(
|
||||
"trail",
|
||||
expect.objectContaining({
|
||||
quoted: expect.objectContaining({
|
||||
key: expect.objectContaining({
|
||||
id: "reply-2",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
}),
|
||||
message: { conversation: "quoted media body" },
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectBuffer(mediaPayload.image, "sendMedia image");
|
||||
expect(mediaPayload.caption).toBe("caption");
|
||||
expect(mediaPayload.mimetype).toBe("image/jpeg");
|
||||
expectQuotedOptions(mockCallArg(msg.sendMedia, 0, 1, "sendMedia"), {
|
||||
id: "reply-2",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
body: "quoted media body",
|
||||
});
|
||||
expect(mockCallArg(msg.reply, 0, 0, "reply")).toBe("trail");
|
||||
expectQuotedOptions(mockCallArg(msg.reply, 0, 1, "reply"), {
|
||||
id: "reply-2",
|
||||
fromMe: true,
|
||||
participant: "111@s.whatsapp.net",
|
||||
body: "quoted media body",
|
||||
});
|
||||
});
|
||||
|
||||
it("retries media send on transient failure", async () => {
|
||||
@@ -566,12 +552,10 @@ describe("deliverWebReply", () => {
|
||||
expect(
|
||||
String((msg.reply as unknown as { mock: { calls: unknown[][] } }).mock.calls[0]?.[0]),
|
||||
).not.toContain("boom");
|
||||
const warnContext = findLoggerContext(
|
||||
replyLogger.warn,
|
||||
expect(replyLogger.warn).toHaveBeenCalledWith(
|
||||
expect.objectContaining({ mediaUrl: "http://example.com/img.jpg" }),
|
||||
"failed to send web media reply",
|
||||
"replyLogger.warn",
|
||||
);
|
||||
expect(warnContext.mediaUrl).toBe("http://example.com/img.jpg");
|
||||
});
|
||||
|
||||
it("still attempts later media after the first media fails", async () => {
|
||||
@@ -618,15 +602,16 @@ describe("deliverWebReply", () => {
|
||||
localRoots: undefined,
|
||||
});
|
||||
expect(msg.sendMedia).toHaveBeenCalledTimes(2);
|
||||
const secondPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 1, 0, "sendMedia"),
|
||||
"second sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
expect.objectContaining({
|
||||
document: expect.any(Buffer),
|
||||
fileName: "good.pdf",
|
||||
caption: undefined,
|
||||
mimetype: "application/pdf",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(secondPayload.document, "second sendMedia document");
|
||||
expect(secondPayload.fileName).toBe("good.pdf");
|
||||
expect(secondPayload.caption).toBeUndefined();
|
||||
expect(secondPayload.mimetype).toBe("application/pdf");
|
||||
expect(mockCallArg(msg.sendMedia, 1, 1, "sendMedia")).toBeUndefined();
|
||||
expect(msg.reply).toHaveBeenCalledTimes(1);
|
||||
expect(
|
||||
String((msg.reply as unknown as { mock: { calls: unknown[][] } }).mock.calls[0]?.[0]),
|
||||
@@ -707,14 +692,14 @@ describe("deliverWebReply", () => {
|
||||
localRoots: undefined,
|
||||
});
|
||||
expect(msg.sendMedia).toHaveBeenCalledTimes(1);
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
audio: expect.any(Buffer),
|
||||
ptt: true,
|
||||
mimetype: "audio/ogg; codecs=opus",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(mediaPayload.audio, "sendMedia audio");
|
||||
expect(mediaPayload.ptt).toBe(true);
|
||||
expect(mediaPayload.mimetype).toBe("audio/ogg; codecs=opus");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
expect(expectFirstSendMediaPayload(msg)).not.toHaveProperty("caption");
|
||||
expect(msg.reply).toHaveBeenCalledWith("caption", undefined);
|
||||
});
|
||||
@@ -738,14 +723,14 @@ describe("deliverWebReply", () => {
|
||||
skipLog: true,
|
||||
});
|
||||
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
audio: expect.any(Buffer),
|
||||
ptt: true,
|
||||
mimetype: "audio/ogg; codecs=opus",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(mediaPayload.audio, "sendMedia audio");
|
||||
expect(mediaPayload.ptt).toBe(true);
|
||||
expect(mediaPayload.mimetype).toBe("audio/ogg; codecs=opus");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
expect(expectFirstSendMediaPayload(msg)).not.toHaveProperty("caption");
|
||||
expect(msg.reply).toHaveBeenCalledWith("cap", undefined);
|
||||
});
|
||||
@@ -775,23 +760,17 @@ describe("deliverWebReply", () => {
|
||||
skipLog: true,
|
||||
});
|
||||
|
||||
const ffmpegArgs = mockCallArg(hoisted.runFfmpeg, 0, 0, "runFfmpeg");
|
||||
expect(Array.isArray(ffmpegArgs)).toBe(true);
|
||||
const ffmpegArgList = ffmpegArgs as unknown[];
|
||||
expect(ffmpegArgList).toContain("-c:a");
|
||||
expect(ffmpegArgList).toContain("libopus");
|
||||
expect(ffmpegArgList).toContain("-ar");
|
||||
expect(ffmpegArgList).toContain("48000");
|
||||
expect(ffmpegArgList).toContain("-b:a");
|
||||
expect(ffmpegArgList).toContain("64k");
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(hoisted.runFfmpeg).toHaveBeenCalledWith(
|
||||
expect.arrayContaining(["-c:a", "libopus", "-ar", "48000", "-b:a", "64k"]),
|
||||
);
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
audio: Buffer.from("opus-output"),
|
||||
ptt: true,
|
||||
mimetype: "audio/ogg; codecs=opus",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expect(mediaPayload.audio).toEqual(Buffer.from("opus-output"));
|
||||
expect(mediaPayload.ptt).toBe(true);
|
||||
expect(mediaPayload.mimetype).toBe("audio/ogg; codecs=opus");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
expect(expectFirstSendMediaPayload(msg)).not.toHaveProperty("caption");
|
||||
expect(msg.reply).toHaveBeenCalledWith("cap", undefined);
|
||||
});
|
||||
@@ -815,14 +794,14 @@ describe("deliverWebReply", () => {
|
||||
skipLog: true,
|
||||
});
|
||||
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
video: expect.any(Buffer),
|
||||
caption: "cap",
|
||||
mimetype: "video/mp4",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(mediaPayload.video, "sendMedia video");
|
||||
expect(mediaPayload.caption).toBe("cap");
|
||||
expect(mediaPayload.mimetype).toBe("video/mp4");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
});
|
||||
|
||||
it("sends non-audio/image/video media as document", async () => {
|
||||
@@ -845,15 +824,15 @@ describe("deliverWebReply", () => {
|
||||
skipLog: true,
|
||||
});
|
||||
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
document: expect.any(Buffer),
|
||||
fileName: "x.bin",
|
||||
caption: "cap",
|
||||
mimetype: "application/octet-stream",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(mediaPayload.document, "sendMedia document");
|
||||
expect(mediaPayload.fileName).toBe("x.bin");
|
||||
expect(mediaPayload.caption).toBe("cap");
|
||||
expect(mediaPayload.mimetype).toBe("application/octet-stream");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
});
|
||||
|
||||
it("strips URL query and fragment data from derived document file names", async () => {
|
||||
@@ -878,14 +857,14 @@ describe("deliverWebReply", () => {
|
||||
skipLog: true,
|
||||
});
|
||||
|
||||
const mediaPayload = requireRecord(
|
||||
mockCallArg(msg.sendMedia, 0, 0, "sendMedia"),
|
||||
"sendMedia payload",
|
||||
expect(msg.sendMedia).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
document: expect.any(Buffer),
|
||||
fileName: "report.pdf",
|
||||
caption: "cap",
|
||||
mimetype: "application/pdf",
|
||||
}),
|
||||
undefined,
|
||||
);
|
||||
expectBuffer(mediaPayload.document, "sendMedia document");
|
||||
expect(mediaPayload.fileName).toBe("report.pdf");
|
||||
expect(mediaPayload.caption).toBe("cap");
|
||||
expect(mediaPayload.mimetype).toBe("application/pdf");
|
||||
expect(mockCallArg(msg.sendMedia, 0, 1, "sendMedia")).toBeUndefined();
|
||||
});
|
||||
});
|
||||
|
||||
@@ -1367,7 +1367,6 @@
|
||||
"crabbox:run": "node scripts/crabbox-wrapper.mjs run",
|
||||
"crabbox:stop": "node scripts/crabbox-wrapper.mjs stop",
|
||||
"crabbox:warmup": "node scripts/crabbox-wrapper.mjs warmup",
|
||||
"discord:opus:install": "node scripts/install-discord-native-opus.mjs",
|
||||
"deadcode:ci": "pnpm deadcode:report:ci:knip && pnpm deadcode:report:ci:ts-unused",
|
||||
"deadcode:dependencies": "pnpm --config.minimum-release-age=0 dlx knip@6.8.0 --config config/knip.config.ts --production --no-progress --reporter compact --dependencies --no-config-hints",
|
||||
"deadcode:knip": "pnpm dlx knip --config config/knip.config.ts --production --no-progress --reporter compact --files --dependencies",
|
||||
|
||||
@@ -1,44 +0,0 @@
|
||||
import { spawnSync } from "node:child_process";
|
||||
import { existsSync } from "node:fs";
|
||||
import path from "node:path";
|
||||
import { fileURLToPath } from "node:url";
|
||||
|
||||
const root = path.resolve(path.dirname(fileURLToPath(import.meta.url)), "..");
|
||||
const opusDir = path.join(root, "node_modules", "@discordjs", "opus");
|
||||
|
||||
if (!existsSync(path.join(opusDir, "package.json"))) {
|
||||
console.error(
|
||||
"Missing node_modules/@discordjs/opus. Run pnpm install first, then retry this opt-in installer.",
|
||||
);
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
const install = spawnSync(
|
||||
"pnpm",
|
||||
["--dir", opusDir, "exec", "node-pre-gyp", "install", "--fallback-to-build"],
|
||||
{
|
||||
cwd: root,
|
||||
env: process.env,
|
||||
stdio: "inherit",
|
||||
},
|
||||
);
|
||||
|
||||
if (install.status !== 0) {
|
||||
console.error(
|
||||
"Failed to install @discordjs/opus for the active Node runtime. Use Node 22 for the upstream macOS arm64 prebuild, or install a node-gyp toolchain for source builds.",
|
||||
);
|
||||
process.exit(install.status ?? 1);
|
||||
}
|
||||
|
||||
const verify = spawnSync(process.execPath, ["-e", 'require("@discordjs/opus")'], {
|
||||
cwd: root,
|
||||
env: process.env,
|
||||
stdio: "inherit",
|
||||
});
|
||||
|
||||
if (verify.status !== 0) {
|
||||
console.error("@discordjs/opus installed, but the active Node runtime still cannot load it.");
|
||||
process.exit(verify.status ?? 1);
|
||||
}
|
||||
|
||||
console.log("native opus ok");
|
||||
@@ -1,548 +0,0 @@
|
||||
import fs from "node:fs/promises";
|
||||
import net from "node:net";
|
||||
import os from "node:os";
|
||||
import path from "node:path";
|
||||
import process from "node:process";
|
||||
import { pathToFileURL } from "node:url";
|
||||
import { startQaMockOpenAiServer } from "../extensions/qa-lab/src/providers/mock-openai/server.js";
|
||||
import { stageQaMockAuthProfiles } from "../extensions/qa-lab/src/providers/shared/mock-auth.js";
|
||||
import { buildQaGatewayConfig } from "../extensions/qa-lab/src/qa-gateway-config.js";
|
||||
import { resetConfigRuntimeState } from "../src/config/config.js";
|
||||
import { startGatewayServer } from "../src/gateway/server.js";
|
||||
|
||||
type Lane = "normal" | "code";
|
||||
|
||||
type LaneResult = {
|
||||
lane: Lane;
|
||||
status: string;
|
||||
providerRequestCount: number;
|
||||
providerRawBytes: number;
|
||||
providerSystemPromptChars: number;
|
||||
providerDeclaredToolCount: number;
|
||||
providerPlannedTools: string[];
|
||||
gatewayOutputToolNames: string[];
|
||||
gatewayOutputText: string;
|
||||
sessionLogToolMentions: Record<string, number>;
|
||||
};
|
||||
|
||||
const FAKE_PLUGIN_ID = "tool-search-e2e-fixture";
|
||||
|
||||
function assert(condition: unknown, message: string): asserts condition {
|
||||
if (!condition) {
|
||||
throw new Error(message);
|
||||
}
|
||||
}
|
||||
|
||||
async function freePort(): Promise<number> {
|
||||
return await new Promise((resolve, reject) => {
|
||||
const server = net.createServer();
|
||||
server.once("error", reject);
|
||||
server.listen(0, "127.0.0.1", () => {
|
||||
const address = server.address();
|
||||
const port = typeof address === "object" && address ? address.port : 0;
|
||||
server.close((error) => (error ? reject(error) : resolve(port)));
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
function buildFakeTools(count = 36) {
|
||||
return Array.from({ length: count }, (_, index) => {
|
||||
const id = `fake_plugin_tool_${String(index + 1).padStart(2, "0")}`;
|
||||
return {
|
||||
type: "function",
|
||||
name: id,
|
||||
description: [
|
||||
`Fake plugin tool ${index + 1}.`,
|
||||
"Used by the Tool Search gateway E2E to prove a large plugin-owned tool catalog can be hidden from the model prompt and still called through the compact bridge.",
|
||||
"The description is intentionally non-trivial so prompt-size regression is measurable.",
|
||||
].join(" "),
|
||||
parameters: {
|
||||
type: "object",
|
||||
properties: {
|
||||
marker: {
|
||||
type: "string",
|
||||
description: "Lane marker supplied by the scripted model.",
|
||||
},
|
||||
},
|
||||
required: ["marker"],
|
||||
additionalProperties: false,
|
||||
},
|
||||
strict: true,
|
||||
};
|
||||
});
|
||||
}
|
||||
|
||||
function countOccurrences(haystack: string, needle: string): number {
|
||||
if (!needle) {
|
||||
return 0;
|
||||
}
|
||||
let count = 0;
|
||||
let offset = 0;
|
||||
while (true) {
|
||||
const next = haystack.indexOf(needle, offset);
|
||||
if (next < 0) {
|
||||
return count;
|
||||
}
|
||||
count += 1;
|
||||
offset = next + needle.length;
|
||||
}
|
||||
}
|
||||
|
||||
async function readSessionLogMentions(params: {
|
||||
stateDir: string;
|
||||
targetTool: string;
|
||||
}): Promise<Record<string, number>> {
|
||||
const sessionsDir = path.join(params.stateDir, "agents", "qa", "sessions");
|
||||
const mentions: Record<string, number> = {
|
||||
tool_search_code: 0,
|
||||
[params.targetTool]: 0,
|
||||
};
|
||||
let files: string[] = [];
|
||||
try {
|
||||
files = await fs.readdir(sessionsDir);
|
||||
} catch {
|
||||
return mentions;
|
||||
}
|
||||
for (const file of files.filter((candidate) => candidate.endsWith(".jsonl"))) {
|
||||
const raw = await fs.readFile(path.join(sessionsDir, file), "utf8").catch(() => "");
|
||||
mentions.tool_search_code += countOccurrences(raw, "tool_search_code");
|
||||
mentions[params.targetTool] += countOccurrences(raw, params.targetTool);
|
||||
}
|
||||
return mentions;
|
||||
}
|
||||
|
||||
async function fetchJson(url: string, init?: RequestInit): Promise<unknown> {
|
||||
const response = await fetch(url, init);
|
||||
const text = await response.text();
|
||||
let parsed: unknown;
|
||||
try {
|
||||
parsed = text ? JSON.parse(text) : {};
|
||||
} catch {
|
||||
parsed = text;
|
||||
}
|
||||
if (!response.ok) {
|
||||
throw new Error(`HTTP ${response.status} from ${url}: ${text}`);
|
||||
}
|
||||
return parsed;
|
||||
}
|
||||
|
||||
function outputToolNames(response: unknown): string[] {
|
||||
const output = (response as { output?: Array<{ type?: unknown; name?: unknown }> }).output;
|
||||
if (!Array.isArray(output)) {
|
||||
return [];
|
||||
}
|
||||
return output
|
||||
.filter((item) => item.type === "function_call" && typeof item.name === "string")
|
||||
.map((item) => item.name as string);
|
||||
}
|
||||
|
||||
function outputText(response: unknown): string {
|
||||
const output = (response as { output?: Array<{ type?: unknown; content?: unknown }> }).output;
|
||||
if (!Array.isArray(output)) {
|
||||
return "";
|
||||
}
|
||||
return output
|
||||
.flatMap((item) => {
|
||||
if (item.type !== "message" || !Array.isArray(item.content)) {
|
||||
return [];
|
||||
}
|
||||
return item.content.flatMap((piece) => {
|
||||
if (!piece || typeof piece !== "object") {
|
||||
return [];
|
||||
}
|
||||
const record = piece as { text?: unknown };
|
||||
return typeof record.text === "string" ? [record.text] : [];
|
||||
});
|
||||
})
|
||||
.join("\n");
|
||||
}
|
||||
|
||||
function readContentText(content: unknown): string {
|
||||
if (typeof content === "string") {
|
||||
return content;
|
||||
}
|
||||
if (!Array.isArray(content)) {
|
||||
return "";
|
||||
}
|
||||
return content
|
||||
.map((item) => {
|
||||
if (!item || typeof item !== "object") {
|
||||
return "";
|
||||
}
|
||||
const record = item as { type?: unknown; text?: unknown };
|
||||
return typeof record.text === "string" ? record.text : "";
|
||||
})
|
||||
.join("\n");
|
||||
}
|
||||
|
||||
function countSystemPromptChars(body: unknown): number {
|
||||
if (!body || typeof body !== "object") {
|
||||
return 0;
|
||||
}
|
||||
const record = body as { instructions?: unknown; input?: unknown };
|
||||
let total = typeof record.instructions === "string" ? record.instructions.length : 0;
|
||||
if (Array.isArray(record.input)) {
|
||||
for (const item of record.input) {
|
||||
if (!item || typeof item !== "object") {
|
||||
continue;
|
||||
}
|
||||
const inputRecord = item as { role?: unknown; content?: unknown };
|
||||
if (inputRecord.role === "system" || inputRecord.role === "developer") {
|
||||
total += readContentText(inputRecord.content).length;
|
||||
}
|
||||
}
|
||||
}
|
||||
return total;
|
||||
}
|
||||
|
||||
async function writeConfig(params: {
|
||||
lane: Lane;
|
||||
stateDir: string;
|
||||
configPath: string;
|
||||
workspaceDir: string;
|
||||
gatewayPort: number;
|
||||
providerBaseUrl: string;
|
||||
fakePluginDir: string;
|
||||
}) {
|
||||
let cfg = buildQaGatewayConfig({
|
||||
bind: "loopback",
|
||||
gatewayPort: params.gatewayPort,
|
||||
gatewayToken: "tool-search-e2e",
|
||||
providerBaseUrl: `${params.providerBaseUrl}/v1`,
|
||||
workspaceDir: params.workspaceDir,
|
||||
controlUiEnabled: false,
|
||||
providerMode: "mock-openai",
|
||||
});
|
||||
cfg = {
|
||||
...cfg,
|
||||
tools: {
|
||||
...cfg.tools,
|
||||
alsoAllow: [...new Set([...(cfg.tools?.alsoAllow ?? []), FAKE_PLUGIN_ID])],
|
||||
},
|
||||
};
|
||||
if (params.lane === "code") {
|
||||
cfg = {
|
||||
...cfg,
|
||||
tools: {
|
||||
...cfg.tools,
|
||||
alsoAllow: [
|
||||
...new Set([
|
||||
...(cfg.tools?.alsoAllow ?? []),
|
||||
"tool_search_code",
|
||||
"tool_search",
|
||||
"tool_describe",
|
||||
"tool_call",
|
||||
]),
|
||||
],
|
||||
toolSearch: true,
|
||||
},
|
||||
plugins: {
|
||||
...cfg.plugins,
|
||||
allow: [...new Set([...(cfg.plugins?.allow ?? []), FAKE_PLUGIN_ID])],
|
||||
entries: {
|
||||
...cfg.plugins?.entries,
|
||||
[FAKE_PLUGIN_ID]: {
|
||||
enabled: true,
|
||||
},
|
||||
},
|
||||
},
|
||||
};
|
||||
} else {
|
||||
cfg = {
|
||||
...cfg,
|
||||
plugins: {
|
||||
...cfg.plugins,
|
||||
allow: [...new Set([...(cfg.plugins?.allow ?? []), FAKE_PLUGIN_ID])],
|
||||
entries: {
|
||||
...cfg.plugins?.entries,
|
||||
[FAKE_PLUGIN_ID]: {
|
||||
enabled: true,
|
||||
},
|
||||
},
|
||||
},
|
||||
};
|
||||
}
|
||||
cfg = {
|
||||
...cfg,
|
||||
plugins: {
|
||||
...cfg.plugins,
|
||||
load: {
|
||||
...cfg.plugins?.load,
|
||||
paths: [...new Set([...(cfg.plugins?.load?.paths ?? []), params.fakePluginDir])],
|
||||
},
|
||||
},
|
||||
};
|
||||
cfg = await stageQaMockAuthProfiles({
|
||||
cfg,
|
||||
stateDir: params.stateDir,
|
||||
agentIds: ["qa"],
|
||||
providers: ["mock-openai", "openai", "anthropic"],
|
||||
});
|
||||
cfg = {
|
||||
...cfg,
|
||||
gateway: {
|
||||
...cfg.gateway,
|
||||
http: {
|
||||
endpoints: {
|
||||
responses: {
|
||||
enabled: true,
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
};
|
||||
await fs.mkdir(path.dirname(params.configPath), { recursive: true });
|
||||
await fs.writeFile(params.configPath, `${JSON.stringify(cfg, null, 2)}\n`, "utf8");
|
||||
}
|
||||
|
||||
async function writeFakePlugin(params: {
|
||||
rootDir: string;
|
||||
repoRoot: string;
|
||||
fakeTools: ReturnType<typeof buildFakeTools>;
|
||||
}): Promise<string> {
|
||||
const pluginDir = path.join(params.rootDir, "fake-plugin");
|
||||
await fs.mkdir(pluginDir, { recursive: true });
|
||||
await fs.writeFile(
|
||||
path.join(pluginDir, "package.json"),
|
||||
`${JSON.stringify(
|
||||
{
|
||||
name: "@openclaw/tool-search-e2e-fixture",
|
||||
version: "0.0.0",
|
||||
type: "module",
|
||||
openclaw: {
|
||||
extensions: ["./index.js"],
|
||||
},
|
||||
},
|
||||
null,
|
||||
2,
|
||||
)}\n`,
|
||||
"utf8",
|
||||
);
|
||||
await fs.writeFile(
|
||||
path.join(pluginDir, "openclaw.plugin.json"),
|
||||
`${JSON.stringify(
|
||||
{
|
||||
id: FAKE_PLUGIN_ID,
|
||||
activation: {
|
||||
onStartup: true,
|
||||
},
|
||||
name: "Tool Search E2E Fixture",
|
||||
description: "Fake plugin with a large tool catalog for Tool Search gateway validation.",
|
||||
contracts: {
|
||||
tools: params.fakeTools.map((tool) => tool.name),
|
||||
},
|
||||
configSchema: {
|
||||
type: "object",
|
||||
additionalProperties: false,
|
||||
properties: {},
|
||||
},
|
||||
},
|
||||
null,
|
||||
2,
|
||||
)}\n`,
|
||||
"utf8",
|
||||
);
|
||||
const pluginEntryUrl = pathToFileURL(
|
||||
path.join(params.repoRoot, "src/plugin-sdk/plugin-entry.ts"),
|
||||
).href;
|
||||
await fs.writeFile(
|
||||
path.join(pluginDir, "index.js"),
|
||||
[
|
||||
`import { definePluginEntry } from ${JSON.stringify(pluginEntryUrl)};`,
|
||||
`const tools = ${JSON.stringify(params.fakeTools, null, 2)};`,
|
||||
"export default definePluginEntry({",
|
||||
` id: ${JSON.stringify(FAKE_PLUGIN_ID)},`,
|
||||
" name: 'Tool Search E2E Fixture',",
|
||||
" register(api) {",
|
||||
" for (const spec of tools) {",
|
||||
" api.registerTool({",
|
||||
" name: spec.name,",
|
||||
" label: spec.name,",
|
||||
" description: spec.description,",
|
||||
" parameters: spec.parameters,",
|
||||
" execute: async (_toolCallId, input) => ({",
|
||||
" content: [{ type: 'text', text: `FAKE_PLUGIN_OK ${spec.name} ${JSON.stringify(input ?? {})}` }],",
|
||||
" details: { status: 'ok', tool: spec.name, input },",
|
||||
" }),",
|
||||
" }, { name: spec.name });",
|
||||
" }",
|
||||
" },",
|
||||
"});",
|
||||
"",
|
||||
].join("\n"),
|
||||
"utf8",
|
||||
);
|
||||
return pluginDir;
|
||||
}
|
||||
|
||||
async function runLane(params: {
|
||||
lane: Lane;
|
||||
rootDir: string;
|
||||
providerBaseUrl: string;
|
||||
targetTool: string;
|
||||
fakeTools: ReturnType<typeof buildFakeTools>;
|
||||
fakePluginDir: string;
|
||||
}): Promise<LaneResult> {
|
||||
const stateDir = path.join(params.rootDir, params.lane, "state");
|
||||
const configPath = path.join(stateDir, "openclaw.json");
|
||||
const workspaceDir = path.join(params.rootDir, params.lane, "workspace");
|
||||
const gatewayPort = await freePort();
|
||||
await fs.mkdir(workspaceDir, { recursive: true });
|
||||
await writeConfig({
|
||||
lane: params.lane,
|
||||
stateDir,
|
||||
configPath,
|
||||
workspaceDir,
|
||||
gatewayPort,
|
||||
providerBaseUrl: params.providerBaseUrl,
|
||||
fakePluginDir: params.fakePluginDir,
|
||||
});
|
||||
|
||||
process.env.OPENCLAW_STATE_DIR = stateDir;
|
||||
process.env.OPENCLAW_CONFIG_PATH = configPath;
|
||||
process.env.OPENCLAW_TEST_FAST = "1";
|
||||
resetConfigRuntimeState();
|
||||
|
||||
const server = await startGatewayServer(gatewayPort, {
|
||||
host: "127.0.0.1",
|
||||
auth: { mode: "none" },
|
||||
controlUiEnabled: false,
|
||||
openResponsesEnabled: true,
|
||||
});
|
||||
try {
|
||||
const beforeRequests = (await fetchJson(
|
||||
`${params.providerBaseUrl}/debug/requests`,
|
||||
)) as unknown[];
|
||||
const response = await fetchJson(`http://127.0.0.1:${gatewayPort}/v1/responses`, {
|
||||
method: "POST",
|
||||
headers: {
|
||||
"content-type": "application/json",
|
||||
"x-openclaw-scopes": "operator.write",
|
||||
"x-openclaw-agent": "qa",
|
||||
},
|
||||
body: JSON.stringify({
|
||||
model: "openclaw/qa",
|
||||
input: [
|
||||
{
|
||||
type: "message",
|
||||
role: "user",
|
||||
content: [
|
||||
{
|
||||
type: "input_text",
|
||||
text: `tool search qa check target=${params.targetTool}`,
|
||||
},
|
||||
],
|
||||
},
|
||||
],
|
||||
max_output_tokens: 256,
|
||||
stream: false,
|
||||
}),
|
||||
});
|
||||
const requests = (await fetchJson(`${params.providerBaseUrl}/debug/requests`)) as Array<{
|
||||
raw?: string;
|
||||
body?: { tools?: unknown[] };
|
||||
instructions?: string;
|
||||
plannedToolName?: string;
|
||||
}>;
|
||||
const laneRequests = requests.slice(beforeRequests.length);
|
||||
const lastRequest = laneRequests.at(-1) ?? {};
|
||||
const responseStatus = (response as { status?: unknown }).status;
|
||||
return {
|
||||
lane: params.lane,
|
||||
status: typeof responseStatus === "string" ? responseStatus : "",
|
||||
providerRequestCount: laneRequests.length,
|
||||
providerRawBytes: typeof lastRequest.raw === "string" ? lastRequest.raw.length : 0,
|
||||
providerSystemPromptChars: countSystemPromptChars(lastRequest.body),
|
||||
providerDeclaredToolCount: Array.isArray(lastRequest.body?.tools)
|
||||
? lastRequest.body.tools.length
|
||||
: 0,
|
||||
providerPlannedTools: laneRequests
|
||||
.map((request) => request.plannedToolName)
|
||||
.filter((name): name is string => typeof name === "string"),
|
||||
gatewayOutputToolNames: outputToolNames(response),
|
||||
gatewayOutputText: outputText(response),
|
||||
sessionLogToolMentions: await readSessionLogMentions({
|
||||
stateDir,
|
||||
targetTool: params.targetTool,
|
||||
}),
|
||||
};
|
||||
} finally {
|
||||
await server.close({ reason: `${params.lane} lane complete` });
|
||||
resetConfigRuntimeState();
|
||||
}
|
||||
}
|
||||
|
||||
async function main() {
|
||||
const rootDir = await fs.mkdtemp(path.join(os.tmpdir(), "openclaw-tool-search-"));
|
||||
const provider = await startQaMockOpenAiServer();
|
||||
const fakeTools = buildFakeTools();
|
||||
const fakePluginDir = await writeFakePlugin({
|
||||
rootDir,
|
||||
repoRoot: process.cwd(),
|
||||
fakeTools,
|
||||
});
|
||||
const targetTool = "fake_plugin_tool_17";
|
||||
try {
|
||||
const normal = await runLane({
|
||||
lane: "normal",
|
||||
rootDir,
|
||||
providerBaseUrl: provider.baseUrl,
|
||||
targetTool,
|
||||
fakeTools,
|
||||
fakePluginDir,
|
||||
});
|
||||
const code = await runLane({
|
||||
lane: "code",
|
||||
rootDir,
|
||||
providerBaseUrl: provider.baseUrl,
|
||||
targetTool,
|
||||
fakeTools,
|
||||
fakePluginDir,
|
||||
});
|
||||
|
||||
assert(
|
||||
normal.providerPlannedTools.includes(targetTool) &&
|
||||
normal.gatewayOutputText.includes("FAKE_PLUGIN_OK") &&
|
||||
normal.gatewayOutputText.includes(targetTool),
|
||||
`normal lane did not call ${targetTool}`,
|
||||
);
|
||||
assert(
|
||||
code.providerPlannedTools.includes("tool_search_code") &&
|
||||
code.gatewayOutputText.includes(targetTool) &&
|
||||
code.sessionLogToolMentions[targetTool] > 0,
|
||||
`code lane did not bridge-call ${targetTool}`,
|
||||
);
|
||||
assert(
|
||||
normal.providerDeclaredToolCount > code.providerDeclaredToolCount,
|
||||
`expected Tool Search to expose fewer tools to provider: normal=${normal.providerDeclaredToolCount} code=${code.providerDeclaredToolCount}`,
|
||||
);
|
||||
assert(
|
||||
normal.providerRawBytes > code.providerRawBytes,
|
||||
`expected Tool Search request to be smaller: normal=${normal.providerRawBytes} code=${code.providerRawBytes}`,
|
||||
);
|
||||
assert(
|
||||
code.sessionLogToolMentions.tool_search_code > 0 &&
|
||||
code.sessionLogToolMentions[targetTool] > 0,
|
||||
"code lane session log did not record bridge and target tool mentions",
|
||||
);
|
||||
|
||||
const summary = {
|
||||
ok: true,
|
||||
rootDir,
|
||||
targetTool,
|
||||
normal,
|
||||
code,
|
||||
reduction: {
|
||||
providerRawBytes: normal.providerRawBytes - code.providerRawBytes,
|
||||
providerDeclaredTools: normal.providerDeclaredToolCount - code.providerDeclaredToolCount,
|
||||
providerSystemPromptChars:
|
||||
normal.providerSystemPromptChars - code.providerSystemPromptChars,
|
||||
},
|
||||
};
|
||||
process.stdout.write(`${JSON.stringify(summary, null, 2)}\n`);
|
||||
} finally {
|
||||
await provider.stop();
|
||||
}
|
||||
}
|
||||
|
||||
await main();
|
||||
@@ -327,70 +327,8 @@ function expectAcceptedSpawn(result: SpawnResult): Extract<SpawnResult, { status
|
||||
return result;
|
||||
}
|
||||
|
||||
function expectRecordFields(
|
||||
record: unknown,
|
||||
expected: Record<string, unknown>,
|
||||
): Record<string, unknown> {
|
||||
expect(record).toBeDefined();
|
||||
const actual = record as Record<string, unknown>;
|
||||
for (const [key, value] of Object.entries(expected)) {
|
||||
expect(actual[key]).toEqual(value);
|
||||
}
|
||||
return actual;
|
||||
}
|
||||
|
||||
function gatewayRequests(): Array<{ method?: string; params?: Record<string, unknown> }> {
|
||||
return hoisted.callGatewayMock.mock.calls.map(
|
||||
(call: unknown[]) => call[0] as { method?: string; params?: Record<string, unknown> },
|
||||
);
|
||||
}
|
||||
|
||||
function gatewayRequest(method: string): { method?: string; params?: Record<string, unknown> } {
|
||||
const request = gatewayRequests().find((candidate) => candidate.method === method);
|
||||
expect(request).toBeDefined();
|
||||
return request as { method?: string; params?: Record<string, unknown> };
|
||||
}
|
||||
|
||||
function expectGatewayMethodNotCalled(method: string): void {
|
||||
expect(gatewayRequests().some((request) => request.method === method)).toBe(false);
|
||||
}
|
||||
|
||||
function expectSessionPatchFields(expected: Record<string, unknown>): void {
|
||||
expectRecordFields(gatewayRequest("sessions.patch").params, expected);
|
||||
}
|
||||
|
||||
function expectInitializeSessionFields(expected: Record<string, unknown>): Record<string, unknown> {
|
||||
return expectRecordFields(hoisted.initializeSessionMock.mock.calls[0]?.[0], expected);
|
||||
}
|
||||
|
||||
function expectBindingCallFields(expected: {
|
||||
conversation?: Record<string, unknown>;
|
||||
metadata?: Record<string, unknown>;
|
||||
placement?: string;
|
||||
targetKind?: string;
|
||||
}): Record<string, unknown> {
|
||||
const input = expectRecordFields(hoisted.sessionBindingBindMock.mock.calls.at(-1)?.[0], {
|
||||
...(expected.placement ? { placement: expected.placement } : {}),
|
||||
...(expected.targetKind ? { targetKind: expected.targetKind } : {}),
|
||||
});
|
||||
if (expected.conversation) {
|
||||
expectRecordFields(input.conversation, expected.conversation);
|
||||
}
|
||||
if (expected.metadata) {
|
||||
expectRecordFields(input.metadata, expected.metadata);
|
||||
}
|
||||
return input;
|
||||
}
|
||||
|
||||
function expectRelayCallFields(expected: Record<string, unknown>, callIndex = 0): void {
|
||||
expectRecordFields(
|
||||
hoisted.startAcpSpawnParentStreamRelayMock.mock.calls[callIndex]?.[0],
|
||||
expected,
|
||||
);
|
||||
}
|
||||
|
||||
function expectAgentGatewayCall(overrides: AgentCallParams): void {
|
||||
const agentCall = gatewayRequest("agent");
|
||||
const agentCall = findAgentGatewayCall();
|
||||
expect(agentCall?.params?.deliver).toBe(overrides.deliver);
|
||||
expect(agentCall?.params?.channel).toBe(overrides.channel);
|
||||
expect(agentCall?.params?.to).toBe(overrides.to);
|
||||
@@ -751,28 +689,37 @@ describe("spawnAcpDirect", () => {
|
||||
expect(accepted.runId).toBe("run-1");
|
||||
expect(accepted.mode).toBe("session");
|
||||
expect(accepted.inlineDelivery).toBe(true);
|
||||
expectSessionPatchFields({
|
||||
const patchCall = hoisted.callGatewayMock.mock.calls
|
||||
.map((call: unknown[]) => call[0] as { method?: string; params?: Record<string, unknown> })
|
||||
.find((request) => request.method === "sessions.patch");
|
||||
expect(patchCall?.params).toMatchObject({
|
||||
key: accepted.childSessionKey,
|
||||
spawnedBy: "agent:main:main",
|
||||
});
|
||||
expectBindingCallFields({
|
||||
targetKind: "session",
|
||||
placement: "child",
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
targetKind: "session",
|
||||
placement: "child",
|
||||
}),
|
||||
);
|
||||
expectResolvedIntroTextInBindMetadata();
|
||||
|
||||
const agentCall = gatewayRequest("agent");
|
||||
const agentCall = hoisted.callGatewayMock.mock.calls
|
||||
.map((call: unknown[]) => call[0] as { method?: string; params?: Record<string, unknown> })
|
||||
.find((request) => request.method === "agent");
|
||||
expect(agentCall?.params?.sessionKey).toMatch(/^agent:codex:acp:/);
|
||||
expect(agentCall?.params?.to).toBe("channel:child-thread");
|
||||
expect(agentCall?.params?.threadId).toBe("child-thread");
|
||||
expect(agentCall?.params?.deliver).toBe(true);
|
||||
expect(agentCall?.params?.lane).toBe("subagent");
|
||||
expect(agentCall?.params?.acpTurnSource).toBe("manual_spawn");
|
||||
const initInput = expectInitializeSessionFields({
|
||||
agent: "codex",
|
||||
mode: "persistent",
|
||||
});
|
||||
expect(initInput.sessionKey).toMatch(/^agent:codex:acp:/);
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: expect.stringMatching(/^agent:codex:acp:/),
|
||||
agent: "codex",
|
||||
mode: "persistent",
|
||||
}),
|
||||
);
|
||||
const transcriptCalls = hoisted.resolveSessionTranscriptFileMock.mock.calls.map(
|
||||
(call: unknown[]) => call[0] as { threadId?: string },
|
||||
);
|
||||
@@ -818,7 +765,11 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expectAcceptedSpawn(result);
|
||||
expectInitializeSessionFields({ resumeSessionId });
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
resumeSessionId,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("rejects ACP resume IDs not recorded for the requester session", async () => {
|
||||
@@ -856,7 +807,7 @@ describe("spawnAcpDirect", () => {
|
||||
},
|
||||
);
|
||||
|
||||
expectRecordFields(result, {
|
||||
expect(result).toMatchObject({
|
||||
status: "forbidden",
|
||||
errorCode: "resume_forbidden",
|
||||
});
|
||||
@@ -878,14 +829,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expectAcceptedSpawn(result);
|
||||
const initInput = expectInitializeSessionFields({
|
||||
agent: "codex",
|
||||
runtimeOptions: {
|
||||
model: "openai-codex/gpt-5.4",
|
||||
thinking: "high",
|
||||
},
|
||||
});
|
||||
expect(initInput.sessionKey).toMatch(/^agent:codex:acp:/);
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: expect.stringMatching(/^agent:codex:acp:/),
|
||||
agent: "codex",
|
||||
runtimeOptions: {
|
||||
model: "openai-codex/gpt-5.4",
|
||||
thinking: "high",
|
||||
},
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("applies ACP spawn run timeout to runtime options and dispatch", async () => {
|
||||
@@ -901,13 +854,15 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expectAcceptedSpawn(result);
|
||||
const initInput = expectInitializeSessionFields({
|
||||
agent: "codex",
|
||||
runtimeOptions: {
|
||||
timeoutSeconds: 45,
|
||||
},
|
||||
});
|
||||
expect(initInput.sessionKey).toMatch(/^agent:codex:acp:/);
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: expect.stringMatching(/^agent:codex:acp:/),
|
||||
agent: "codex",
|
||||
runtimeOptions: {
|
||||
timeoutSeconds: 45,
|
||||
},
|
||||
}),
|
||||
);
|
||||
const agentCall = findAgentGatewayCall();
|
||||
expect(agentCall?.params?.lane).toBe("subagent");
|
||||
expect(agentCall?.params?.timeout).toBe(45);
|
||||
@@ -942,13 +897,15 @@ describe("spawnAcpDirect", () => {
|
||||
},
|
||||
);
|
||||
|
||||
expectRecordFields(result, {
|
||||
expect(result).toMatchObject({
|
||||
status: "error",
|
||||
errorCode: "runtime_agent_mismatch",
|
||||
});
|
||||
expect(result).toHaveProperty("error", expect.stringContaining("OpenClaw config agent"));
|
||||
expect(hoisted.initializeSessionMock).not.toHaveBeenCalled();
|
||||
expectGatewayMethodNotCalled("agent");
|
||||
expect(hoisted.callGatewayMock).not.toHaveBeenCalledWith(
|
||||
expect.objectContaining({ method: "agent" }),
|
||||
);
|
||||
});
|
||||
|
||||
it("maps OpenClaw ACP runtime agent aliases to their configured harness id", async () => {
|
||||
@@ -986,8 +943,12 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expectAcceptedSpawn(result);
|
||||
const initInput = expectInitializeSessionFields({ agent: "codex" });
|
||||
expect(initInput.sessionKey).toMatch(/^agent:codex:acp:/);
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
agent: "codex",
|
||||
sessionKey: expect.stringMatching(/^agent:codex:acp:/),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("inherits subagent envelope fields onto ACP children", async () => {
|
||||
@@ -1010,7 +971,10 @@ describe("spawnAcpDirect", () => {
|
||||
});
|
||||
|
||||
const accepted = expectAcceptedSpawn(result);
|
||||
expectSessionPatchFields({
|
||||
const patchCall = hoisted.callGatewayMock.mock.calls
|
||||
.map((call: unknown[]) => call[0] as { method?: string; params?: Record<string, unknown> })
|
||||
.find((request) => request.method === "sessions.patch");
|
||||
expect(patchCall?.params).toMatchObject({
|
||||
key: accepted.childSessionKey,
|
||||
spawnedBy: "agent:main:subagent:parent",
|
||||
spawnDepth: 2,
|
||||
@@ -1297,14 +1261,16 @@ describe("spawnAcpDirect", () => {
|
||||
},
|
||||
);
|
||||
expect(result.status, JSON.stringify(result)).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "child",
|
||||
conversation: {
|
||||
channel: "matrix",
|
||||
accountId: "default",
|
||||
conversationId: "!room:example",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "child",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "matrix",
|
||||
accountId: "default",
|
||||
conversationId: "!room:example",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectAgentGatewayCall({
|
||||
deliver: true,
|
||||
channel: "matrix",
|
||||
@@ -1355,14 +1321,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status, JSON.stringify(result)).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "child",
|
||||
conversation: {
|
||||
channel: "matrix",
|
||||
accountId: "default",
|
||||
conversationId: "!Room:Example.org",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "child",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "matrix",
|
||||
accountId: "default",
|
||||
conversationId: "!Room:Example.org",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectAgentGatewayCall({
|
||||
deliver: true,
|
||||
channel: "matrix",
|
||||
@@ -1414,15 +1382,17 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status, JSON.stringify(result)).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "child",
|
||||
conversation: {
|
||||
channel: "matrix",
|
||||
accountId: "default",
|
||||
conversationId: "$thread-root",
|
||||
parentConversationId: "!Room:Example.org",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "child",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "matrix",
|
||||
accountId: "default",
|
||||
conversationId: "$thread-root",
|
||||
parentConversationId: "!Room:Example.org",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectAgentGatewayCall({
|
||||
deliver: true,
|
||||
channel: "matrix",
|
||||
@@ -1448,11 +1418,13 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
const initInput = expectInitializeSessionFields({
|
||||
agent: "claude-code",
|
||||
cwd: fixture.targetWorkspace,
|
||||
});
|
||||
expect(initInput.sessionKey).toMatch(/^agent:claude-code:acp:/);
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: expect.stringMatching(/^agent:claude-code:acp:/),
|
||||
agent: "claude-code",
|
||||
cwd: fixture.targetWorkspace,
|
||||
}),
|
||||
);
|
||||
} finally {
|
||||
await fs.rm(fixture.workspaceRoot, { recursive: true, force: true });
|
||||
}
|
||||
@@ -1478,11 +1450,13 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
const initInput = expectInitializeSessionFields({
|
||||
agent: "claude-code",
|
||||
cwd: undefined,
|
||||
});
|
||||
expect(initInput.sessionKey).toMatch(/^agent:claude-code:acp:/);
|
||||
expect(hoisted.initializeSessionMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionKey: expect.stringMatching(/^agent:claude-code:acp:/),
|
||||
agent: "claude-code",
|
||||
cwd: undefined,
|
||||
}),
|
||||
);
|
||||
} finally {
|
||||
await fs.rm(fixture.workspaceRoot, { recursive: true, force: true });
|
||||
}
|
||||
@@ -1560,14 +1534,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status, JSON.stringify(result)).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "line",
|
||||
accountId: "default",
|
||||
conversationId: "U1234567890abcdef1234567890abcdef",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "line",
|
||||
accountId: "default",
|
||||
conversationId: "U1234567890abcdef1234567890abcdef",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectAgentGatewayCall({
|
||||
deliver: true,
|
||||
channel: "line",
|
||||
@@ -1653,14 +1629,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "custom",
|
||||
accountId: "work",
|
||||
conversationId: "123456",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "custom",
|
||||
accountId: "work",
|
||||
conversationId: "123456",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectAgentGatewayCall({
|
||||
deliver: true,
|
||||
channel: "custom",
|
||||
@@ -1762,15 +1740,17 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "child",
|
||||
conversation: {
|
||||
channel: "matrix",
|
||||
accountId: "bot-alpha",
|
||||
conversationId: boundRoom,
|
||||
},
|
||||
});
|
||||
expectRecordFields(gatewayRequest("agent").params, {
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "child",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "matrix",
|
||||
accountId: "bot-alpha",
|
||||
conversationId: boundRoom,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(findAgentGatewayCall()?.params).toMatchObject({
|
||||
deliver: true,
|
||||
channel: "matrix",
|
||||
accountId: "bot-alpha",
|
||||
@@ -1840,14 +1820,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "line",
|
||||
accountId: "default",
|
||||
conversationId: expectedConversationId,
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "line",
|
||||
accountId: "default",
|
||||
conversationId: expectedConversationId,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
},
|
||||
);
|
||||
|
||||
@@ -1891,14 +1873,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "line",
|
||||
accountId: "default",
|
||||
conversationId: "R1234567890abcdef1234567890abcdef",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "line",
|
||||
accountId: "default",
|
||||
conversationId: "R1234567890abcdef1234567890abcdef",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it.each([
|
||||
@@ -1935,11 +1919,13 @@ describe("spawnAcpDirect", () => {
|
||||
expect(accepted.streamLogPath).toBeUndefined();
|
||||
expect(hoisted.startAcpSpawnParentStreamRelayMock).not.toHaveBeenCalled();
|
||||
if (expectTranscriptPersistence) {
|
||||
expectRecordFields(hoisted.resolveSessionTranscriptFileMock.mock.calls[0]?.[0], {
|
||||
sessionId: "sess-123",
|
||||
storePath: "/tmp/codex-sessions.json",
|
||||
agentId: "codex",
|
||||
});
|
||||
expect(hoisted.resolveSessionTranscriptFileMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
sessionId: "sess-123",
|
||||
storePath: "/tmp/codex-sessions.json",
|
||||
agentId: "codex",
|
||||
}),
|
||||
);
|
||||
}
|
||||
expectAgentGatewayCall(expectedAgentCall);
|
||||
});
|
||||
@@ -1988,10 +1974,13 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
const bindInput = expectBindingCallFields({});
|
||||
const metadata = expectRecordFields(bindInput.metadata, {});
|
||||
expect(typeof metadata.introText).toBe("string");
|
||||
expect(metadata.introText).toContain("cwd: /home/bob/clawd");
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
metadata: expect.objectContaining({
|
||||
introText: expect.stringContaining("cwd: /home/bob/clawd"),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("rejects disallowed ACP agents", async () => {
|
||||
@@ -2014,7 +2003,7 @@ describe("spawnAcpDirect", () => {
|
||||
},
|
||||
);
|
||||
|
||||
expectRecordFields(result, {
|
||||
expect(result).toMatchObject({
|
||||
status: "forbidden",
|
||||
});
|
||||
});
|
||||
@@ -2143,22 +2132,22 @@ describe("spawnAcpDirect", () => {
|
||||
expect(typeof relayCallOrder).toBe("number");
|
||||
expect(typeof agentCallOrder).toBe("number");
|
||||
expect(relayCallOrder < agentCallOrder).toBe(true);
|
||||
expectRelayCallFields({
|
||||
parentSessionKey: "agent:main:main",
|
||||
agentId: "codex",
|
||||
logPath: "/tmp/sess-main.acp-stream.jsonl",
|
||||
emitStartNotice: false,
|
||||
});
|
||||
expect(hoisted.startAcpSpawnParentStreamRelayMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
parentSessionKey: "agent:main:main",
|
||||
agentId: "codex",
|
||||
logPath: "/tmp/sess-main.acp-stream.jsonl",
|
||||
emitStartNotice: false,
|
||||
}),
|
||||
);
|
||||
const relayRuns = hoisted.startAcpSpawnParentStreamRelayMock.mock.calls.map(
|
||||
(call: unknown[]) => (call[0] as { runId?: string }).runId,
|
||||
);
|
||||
expect(relayRuns).toContain(agentCall?.params?.idempotencyKey);
|
||||
expect(relayRuns).toContain(accepted.runId);
|
||||
const streamPathInput = expectRecordFields(
|
||||
hoisted.resolveAcpSpawnStreamLogPathMock.mock.calls[0]?.[0],
|
||||
{},
|
||||
);
|
||||
expect(streamPathInput.childSessionKey).toMatch(/^agent:codex:acp:/);
|
||||
expect(hoisted.resolveAcpSpawnStreamLogPathMock).toHaveBeenCalledWith({
|
||||
childSessionKey: expect.stringMatching(/^agent:codex:acp:/),
|
||||
});
|
||||
expect(firstHandle.dispose).toHaveBeenCalledTimes(1);
|
||||
expect(firstHandle.notifyStarted).not.toHaveBeenCalled();
|
||||
expect(secondHandle.notifyStarted).toHaveBeenCalledTimes(1);
|
||||
@@ -2231,17 +2220,19 @@ describe("spawnAcpDirect", () => {
|
||||
expect(agentCall?.params?.channel).toBeUndefined();
|
||||
expect(agentCall?.params?.to).toBeUndefined();
|
||||
expect(agentCall?.params?.threadId).toBeUndefined();
|
||||
expectRelayCallFields({
|
||||
parentSessionKey: "agent:main:subagent:parent",
|
||||
agentId: "codex",
|
||||
logPath: "/tmp/sess-main.acp-stream.jsonl",
|
||||
deliveryContext: {
|
||||
channel: "discord",
|
||||
to: "channel:parent-channel",
|
||||
accountId: "default",
|
||||
},
|
||||
emitStartNotice: false,
|
||||
});
|
||||
expect(hoisted.startAcpSpawnParentStreamRelayMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
parentSessionKey: "agent:main:subagent:parent",
|
||||
agentId: "codex",
|
||||
logPath: "/tmp/sess-main.acp-stream.jsonl",
|
||||
deliveryContext: {
|
||||
channel: "discord",
|
||||
to: "channel:parent-channel",
|
||||
accountId: "default",
|
||||
},
|
||||
emitStartNotice: false,
|
||||
}),
|
||||
);
|
||||
expect(firstHandle.dispose).toHaveBeenCalledTimes(1);
|
||||
expect(secondHandle.notifyStarted).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
@@ -2579,15 +2570,17 @@ describe("spawnAcpDirect", () => {
|
||||
|
||||
const accepted = expectAcceptedSpawn(result);
|
||||
expect(accepted.mode).toBe("session");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "telegram",
|
||||
accountId: "default",
|
||||
conversationId: "2",
|
||||
parentConversationId: "-1003342490704",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "telegram",
|
||||
accountId: "default",
|
||||
conversationId: "2",
|
||||
parentConversationId: "-1003342490704",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const agentCall = hoisted.callGatewayMock.mock.calls
|
||||
.map((call: unknown[]) => call[0] as { method?: string; params?: Record<string, unknown> })
|
||||
.find((request) => request.method === "agent");
|
||||
@@ -2615,14 +2608,16 @@ describe("spawnAcpDirect", () => {
|
||||
|
||||
const accepted = expectAcceptedSpawn(result);
|
||||
expect(accepted.mode).toBe("session");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "telegram",
|
||||
accountId: "default",
|
||||
conversationId: "6098642967",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "telegram",
|
||||
accountId: "default",
|
||||
conversationId: "6098642967",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const bindCall = hoisted.sessionBindingBindMock.mock.calls.at(-1)?.[0] as
|
||||
| { conversation?: { parentConversationId?: string } }
|
||||
| undefined;
|
||||
@@ -2648,14 +2643,16 @@ describe("spawnAcpDirect", () => {
|
||||
);
|
||||
|
||||
expect(result.status).toBe("accepted");
|
||||
expectBindingCallFields({
|
||||
placement: "current",
|
||||
conversation: {
|
||||
channel: "telegram",
|
||||
accountId: "default",
|
||||
conversationId: "-1003342490704:topic:2",
|
||||
},
|
||||
});
|
||||
expect(hoisted.sessionBindingBindMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
placement: "current",
|
||||
conversation: expect.objectContaining({
|
||||
channel: "telegram",
|
||||
accountId: "default",
|
||||
conversationId: "-1003342490704:topic:2",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("disposes pre-registered parent relay when initial ACP dispatch fails", async () => {
|
||||
|
||||
@@ -74,30 +74,6 @@ function latestAnthropicRequestHeaders() {
|
||||
return new Headers(latestAnthropicRequest().init?.headers);
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): unknown[] {
|
||||
if (!Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value;
|
||||
}
|
||||
|
||||
function findRecord(items: unknown, predicate: (record: Record<string, unknown>) => boolean) {
|
||||
for (const item of requireArray(items, "items")) {
|
||||
const record = requireRecord(item, "item");
|
||||
if (predicate(record)) {
|
||||
return record;
|
||||
}
|
||||
}
|
||||
throw new Error("Expected matching record");
|
||||
}
|
||||
|
||||
function makeAnthropicTransportModel(
|
||||
params: {
|
||||
id?: string;
|
||||
@@ -178,19 +154,25 @@ describe("anthropic transport stream", () => {
|
||||
);
|
||||
|
||||
expect(buildGuardedModelFetchMock).toHaveBeenCalledWith(model);
|
||||
const [url, init] = guardedFetchMock.mock.calls[0] ?? [];
|
||||
expect(url).toBe("https://api.anthropic.com/v1/messages");
|
||||
expect(init?.method).toBe("POST");
|
||||
const headers = new Headers(init?.headers);
|
||||
expect(headers.get("x-api-key")).toBe("sk-ant-api");
|
||||
expect(headers.get("anthropic-version")).toBe("2023-06-01");
|
||||
expect(headers.get("content-type")).toBe("application/json");
|
||||
expect(headers.get("accept")).toBe("application/json");
|
||||
expect(headers.get("anthropic-dangerous-direct-browser-access")).toBe("true");
|
||||
expect(headers.get("X-Provider")).toBe("anthropic");
|
||||
expect(headers.get("X-Call")).toBe("1");
|
||||
expect(latestAnthropicRequest().payload.model).toBe("claude-sonnet-4-6");
|
||||
expect(latestAnthropicRequest().payload.stream).toBe(true);
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
"https://api.anthropic.com/v1/messages",
|
||||
expect.objectContaining({
|
||||
method: "POST",
|
||||
headers: expect.objectContaining({
|
||||
"x-api-key": "sk-ant-api",
|
||||
"anthropic-version": "2023-06-01",
|
||||
"content-type": "application/json",
|
||||
accept: "application/json",
|
||||
"anthropic-dangerous-direct-browser-access": "true",
|
||||
"X-Provider": "anthropic",
|
||||
"X-Call": "1",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(latestAnthropicRequest().payload).toMatchObject({
|
||||
model: "claude-sonnet-4-6",
|
||||
stream: true,
|
||||
});
|
||||
expect(latestAnthropicRequestHeaders().get("anthropic-beta")).toBe(
|
||||
"fine-grained-tool-streaming-2025-05-14",
|
||||
);
|
||||
@@ -212,8 +194,10 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
expect(guardedFetchMock.mock.calls[0]?.[0]).toBe("https://custom-proxy.example/v1/messages");
|
||||
expect(guardedFetchMock.mock.calls[0]?.[1]?.method).toBe("POST");
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
"https://custom-proxy.example/v1/messages",
|
||||
expect.objectContaining({ method: "POST" }),
|
||||
);
|
||||
expect(latestAnthropicRequestHeaders().get("anthropic-beta")).toBeNull();
|
||||
});
|
||||
|
||||
@@ -294,9 +278,11 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
expect(latestAnthropicRequest().payload.model).toBe("claude-sonnet-4-6");
|
||||
expect(latestAnthropicRequest().payload.max_tokens).toBe(8192);
|
||||
expect(latestAnthropicRequest().payload.stream).toBe(true);
|
||||
expect(latestAnthropicRequest().payload).toMatchObject({
|
||||
model: "claude-sonnet-4-6",
|
||||
max_tokens: 8192,
|
||||
stream: true,
|
||||
});
|
||||
});
|
||||
|
||||
it("ignores fractional runtime maxTokens overrides that floor to zero", async () => {
|
||||
@@ -311,9 +297,11 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
expect(latestAnthropicRequest().payload.model).toBe("claude-sonnet-4-6");
|
||||
expect(latestAnthropicRequest().payload.max_tokens).toBe(8192);
|
||||
expect(latestAnthropicRequest().payload.stream).toBe(true);
|
||||
expect(latestAnthropicRequest().payload).toMatchObject({
|
||||
model: "claude-sonnet-4-6",
|
||||
max_tokens: 8192,
|
||||
stream: true,
|
||||
});
|
||||
});
|
||||
|
||||
it("fails locally when Anthropic maxTokens is non-positive after resolution", async () => {
|
||||
@@ -439,31 +427,33 @@ describe("anthropic transport stream", () => {
|
||||
);
|
||||
const result = await stream.result();
|
||||
|
||||
expect(guardedFetchMock.mock.calls[0]?.[0]).toBe("https://api.anthropic.com/v1/messages");
|
||||
const headers = new Headers(guardedFetchMock.mock.calls[0]?.[1]?.headers);
|
||||
expect(headers.get("authorization")).toBe("Bearer sk-ant-oat-example");
|
||||
expect(headers.get("x-app")).toBe("cli");
|
||||
expect(headers.get("user-agent")).toContain("claude-cli/");
|
||||
expect(guardedFetchMock).toHaveBeenCalledWith(
|
||||
"https://api.anthropic.com/v1/messages",
|
||||
expect.objectContaining({
|
||||
headers: expect.objectContaining({
|
||||
authorization: "Bearer sk-ant-oat-example",
|
||||
"x-app": "cli",
|
||||
"user-agent": expect.stringContaining("claude-cli/"),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const firstCallParams = latestAnthropicRequest().payload;
|
||||
const system = requireArray(firstCallParams.system, "system");
|
||||
expect(
|
||||
system.some(
|
||||
(item) =>
|
||||
requireRecord(item, "system item").text ===
|
||||
"You are Claude Code, Anthropic's official CLI for Claude.",
|
||||
),
|
||||
).toBe(true);
|
||||
expect(
|
||||
system.some((item) => requireRecord(item, "system item").text === "Follow policy."),
|
||||
).toBe(true);
|
||||
expect(
|
||||
requireArray(firstCallParams.tools, "tools").some(
|
||||
(item) => requireRecord(item, "tool").name === "Read",
|
||||
),
|
||||
).toBe(true);
|
||||
expect(firstCallParams.system).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
text: "You are Claude Code, Anthropic's official CLI for Claude.",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
text: "Follow policy.",
|
||||
}),
|
||||
]),
|
||||
);
|
||||
expect(firstCallParams.tools).toEqual(
|
||||
expect.arrayContaining([expect.objectContaining({ name: "Read" })]),
|
||||
);
|
||||
expect(result.stopReason).toBe("toolUse");
|
||||
expect(result.content.some((item) => item.type === "toolCall" && item.name === "read")).toBe(
|
||||
true,
|
||||
expect(result.content).toEqual(
|
||||
expect.arrayContaining([expect.objectContaining({ type: "toolCall", name: "read" })]),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -526,16 +516,19 @@ describe("anthropic transport stream", () => {
|
||||
}
|
||||
const result = await stream.result();
|
||||
|
||||
const thinkingContent = requireRecord(result.content[0], "thinking content");
|
||||
expect(thinkingContent.type).toBe("thinking");
|
||||
expect(thinkingContent.thinking).toBe("checking");
|
||||
expect(thinkingContent.thinkingSignature).toBe("sig_2");
|
||||
expect(result.content[1]).toEqual({ type: "text", text: "NO_REPLY" });
|
||||
expect(events.some((event) => event.type === "text_delta" && event.delta === "NO_REPLY")).toBe(
|
||||
true,
|
||||
);
|
||||
expect(events.some((event) => event.type === "text_end" && event.content === "NO_REPLY")).toBe(
|
||||
true,
|
||||
expect(result.content).toEqual([
|
||||
expect.objectContaining({
|
||||
type: "thinking",
|
||||
thinking: "checking",
|
||||
thinkingSignature: "sig_2",
|
||||
}),
|
||||
{ type: "text", text: "NO_REPLY" },
|
||||
]);
|
||||
expect(events).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ type: "text_delta", delta: "NO_REPLY" }),
|
||||
expect.objectContaining({ type: "text_end", content: "NO_REPLY" }),
|
||||
]),
|
||||
);
|
||||
expect(result.usage.output).toBe(9);
|
||||
});
|
||||
@@ -590,12 +583,12 @@ describe("anthropic transport stream", () => {
|
||||
|
||||
expect(result.content).toEqual([{ type: "text", text: "你好" }]);
|
||||
expect(result.stopReason).toBe("stop");
|
||||
expect(events.some((event) => event.type === "text_start")).toBe(true);
|
||||
expect(events.some((event) => event.type === "text_delta" && event.delta === "你好")).toBe(
|
||||
true,
|
||||
);
|
||||
expect(events.some((event) => event.type === "text_end" && event.content === "你好")).toBe(
|
||||
true,
|
||||
expect(events).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ type: "text_start" }),
|
||||
expect.objectContaining({ type: "text_delta", delta: "你好" }),
|
||||
expect.objectContaining({ type: "text_end", content: "你好" }),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -628,13 +621,16 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
const tools = requireArray(latestAnthropicRequest().payload.tools, "tools");
|
||||
expect(tools).toHaveLength(1);
|
||||
const tool = requireRecord(tools[0], "tool");
|
||||
expect(tool.name).toBe("good_plugin_tool");
|
||||
expect(requireRecord(tool.input_schema, "input schema").properties).toEqual({
|
||||
query: { type: "string" },
|
||||
});
|
||||
expect(latestAnthropicRequest().payload.tools).toEqual([
|
||||
expect.objectContaining({
|
||||
name: "good_plugin_tool",
|
||||
input_schema: expect.objectContaining({
|
||||
properties: {
|
||||
query: { type: "string" },
|
||||
},
|
||||
}),
|
||||
}),
|
||||
]);
|
||||
});
|
||||
|
||||
it("coerces replayed malformed tool-call args to an object for Anthropic payloads", async () => {
|
||||
@@ -678,15 +674,20 @@ describe("anthropic transport stream", () => {
|
||||
await stream.result();
|
||||
|
||||
const firstCallParams = latestAnthropicRequest().payload;
|
||||
const assistantMessage = findRecord(
|
||||
firstCallParams.messages,
|
||||
(record) => record.role === "assistant",
|
||||
expect(firstCallParams.messages).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "assistant",
|
||||
content: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
type: "tool_use",
|
||||
name: "lookup",
|
||||
input: {},
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
]),
|
||||
);
|
||||
const toolUse = findRecord(
|
||||
assistantMessage.content,
|
||||
(record) => record.type === "tool_use" && record.name === "lookup",
|
||||
);
|
||||
expect(toolUse.input).toEqual({});
|
||||
});
|
||||
|
||||
it.each([
|
||||
@@ -720,16 +721,21 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
const userMessage = findRecord(
|
||||
latestAnthropicRequest().payload.messages,
|
||||
(record) => record.role === "user",
|
||||
expect(latestAnthropicRequest().payload.messages).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
type: "tool_result",
|
||||
tool_use_id: "tool_1",
|
||||
content: "(no output)",
|
||||
is_error: false,
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
]),
|
||||
);
|
||||
const toolResult = findRecord(
|
||||
userMessage.content,
|
||||
(record) => record.type === "tool_result" && record.tool_use_id === "tool_1",
|
||||
);
|
||||
expect(toolResult.content).toBe("(no output)");
|
||||
expect(toolResult.is_error).toBe(false);
|
||||
});
|
||||
|
||||
it("drops empty text blocks from image tool results before Anthropic payloads", async () => {
|
||||
@@ -764,26 +770,31 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
const userMessage = findRecord(
|
||||
latestAnthropicRequest().payload.messages,
|
||||
(record) => record.role === "user",
|
||||
expect(latestAnthropicRequest().payload.messages).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
type: "tool_result",
|
||||
tool_use_id: "tool_1",
|
||||
content: [
|
||||
{ type: "text", text: "(see attached image)" },
|
||||
{
|
||||
type: "image",
|
||||
source: {
|
||||
type: "base64",
|
||||
media_type: "image/png",
|
||||
data: imageData,
|
||||
},
|
||||
},
|
||||
],
|
||||
is_error: false,
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
]),
|
||||
);
|
||||
const toolResult = findRecord(
|
||||
userMessage.content,
|
||||
(record) => record.type === "tool_result" && record.tool_use_id === "tool_1",
|
||||
);
|
||||
expect(toolResult.content).toEqual([
|
||||
{ type: "text", text: "(see attached image)" },
|
||||
{
|
||||
type: "image",
|
||||
source: {
|
||||
type: "base64",
|
||||
media_type: "image/png",
|
||||
data: imageData,
|
||||
},
|
||||
},
|
||||
]);
|
||||
expect(toolResult.is_error).toBe(false);
|
||||
});
|
||||
|
||||
it("cancels stalled SSE body reads when the abort signal fires mid-stream", async () => {
|
||||
@@ -862,9 +873,10 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
const payload = latestAnthropicRequest().payload;
|
||||
expect(payload.thinking).toEqual({ type: "adaptive" });
|
||||
expect(payload.output_config).toEqual({ effort: "max" });
|
||||
expect(latestAnthropicRequest().payload).toMatchObject({
|
||||
thinking: { type: "adaptive" },
|
||||
output_config: { effort: "max" },
|
||||
});
|
||||
});
|
||||
|
||||
it("maps xhigh thinking effort for Claude Opus 4.7 transport runs", async () => {
|
||||
@@ -885,8 +897,9 @@ describe("anthropic transport stream", () => {
|
||||
} as AnthropicStreamOptions,
|
||||
);
|
||||
|
||||
const payload = latestAnthropicRequest().payload;
|
||||
expect(payload.thinking).toEqual({ type: "adaptive" });
|
||||
expect(payload.output_config).toEqual({ effort: "xhigh" });
|
||||
expect(latestAnthropicRequest().payload).toMatchObject({
|
||||
thinking: { type: "adaptive" },
|
||||
output_config: { effort: "xhigh" },
|
||||
});
|
||||
});
|
||||
});
|
||||
|
||||
@@ -25,9 +25,6 @@ export type SessionStdin = {
|
||||
// When backed by a real Node stream (child.stdin), this exists; for PTY wrappers it may not.
|
||||
destroy?: () => void;
|
||||
destroyed?: boolean;
|
||||
writable?: boolean;
|
||||
writableEnded?: boolean;
|
||||
writableFinished?: boolean;
|
||||
};
|
||||
|
||||
export interface ProcessSession {
|
||||
|
||||
@@ -66,7 +66,7 @@ export function describeExecTool(params?: { agentId?: string; hasCronTool?: bool
|
||||
export function describeProcessTool(params?: { hasCronTool?: boolean }): string {
|
||||
return [
|
||||
"Manage running exec sessions for commands already started: list, poll, log, write, send-keys, submit, paste, kill.",
|
||||
"Use poll/log when you need status, logs, quiet-success confirmation, or completion confirmation when automatic completion wake is unavailable. Use poll/log also for input-wait hints. Use write/send-keys/submit/paste/kill for input or intervention.",
|
||||
"Use poll/log when you need status, logs, quiet-success confirmation, or completion confirmation when automatic completion wake is unavailable. Use write/send-keys/submit/paste/kill for input or intervention.",
|
||||
params?.hasCronTool
|
||||
? "Do not use process polling to emulate timers or reminders; use cron for scheduled follow-ups."
|
||||
: undefined,
|
||||
|
||||
@@ -3,30 +3,16 @@ import { normalizeOptionalString } from "../shared/string-coerce.js";
|
||||
import type { ExecElevatedDefaults } from "./bash-tools.exec-types.js";
|
||||
|
||||
const EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_PREFIX = "exec-approval-followup:";
|
||||
const EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_NONCE_MARKER = ":nonce:";
|
||||
const EXEC_APPROVAL_FOLLOWUP_RUNTIME_HANDOFF_TTL_MS = 5 * 60 * 1000;
|
||||
const EXEC_APPROVAL_FOLLOWUP_ELEVATED_TOKEN_MARKER = ":elevated:";
|
||||
const EXEC_APPROVAL_FOLLOWUP_ELEVATED_TTL_MS = 5 * 60 * 1000;
|
||||
|
||||
export type ExecApprovalFollowupRuntimeHandoff = {
|
||||
kind: "exec-approval-followup";
|
||||
approvalId: string;
|
||||
type ExecApprovalFollowupElevatedEntry = {
|
||||
sessionKey: string;
|
||||
idempotencyKey: string;
|
||||
bashElevated: ExecElevatedDefaults;
|
||||
};
|
||||
|
||||
export type ExecApprovalFollowupRuntimeHandoffRegistration = {
|
||||
handoffId: string;
|
||||
idempotencyKey: string;
|
||||
};
|
||||
|
||||
type ExecApprovalFollowupRuntimeHandoffEntry = ExecApprovalFollowupRuntimeHandoff & {
|
||||
expiresAtMs: number;
|
||||
};
|
||||
|
||||
const execApprovalFollowupRuntimeHandoffs = new Map<
|
||||
string,
|
||||
ExecApprovalFollowupRuntimeHandoffEntry
|
||||
>();
|
||||
const execApprovalFollowupElevatedDefaults = new Map<string, ExecApprovalFollowupElevatedEntry>();
|
||||
|
||||
function cloneExecElevatedDefaults(value: ExecElevatedDefaults): ExecElevatedDefaults {
|
||||
return {
|
||||
@@ -42,109 +28,96 @@ function cloneExecElevatedDefaults(value: ExecElevatedDefaults): ExecElevatedDef
|
||||
};
|
||||
}
|
||||
|
||||
function cloneExecApprovalFollowupRuntimeHandoff(
|
||||
value: ExecApprovalFollowupRuntimeHandoff,
|
||||
): ExecApprovalFollowupRuntimeHandoff {
|
||||
return {
|
||||
kind: value.kind,
|
||||
approvalId: value.approvalId,
|
||||
sessionKey: value.sessionKey,
|
||||
idempotencyKey: value.idempotencyKey,
|
||||
bashElevated: cloneExecElevatedDefaults(value.bashElevated),
|
||||
};
|
||||
}
|
||||
|
||||
function pruneExpiredExecApprovalFollowupRuntimeHandoffs(nowMs: number): void {
|
||||
for (const [handoffId, entry] of execApprovalFollowupRuntimeHandoffs) {
|
||||
function pruneExpiredExecApprovalFollowupElevatedDefaults(nowMs: number): void {
|
||||
for (const [token, entry] of execApprovalFollowupElevatedDefaults) {
|
||||
if (entry.expiresAtMs <= nowMs) {
|
||||
execApprovalFollowupRuntimeHandoffs.delete(handoffId);
|
||||
execApprovalFollowupElevatedDefaults.delete(token);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
export function buildExecApprovalFollowupIdempotencyKey(params: {
|
||||
approvalId: string;
|
||||
nonce?: string;
|
||||
execApprovalFollowupToken?: string;
|
||||
}): string {
|
||||
const base = `${EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_PREFIX}${params.approvalId}`;
|
||||
const nonce = normalizeOptionalString(params.nonce);
|
||||
return nonce ? `${base}${EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_NONCE_MARKER}${nonce}` : base;
|
||||
return params.execApprovalFollowupToken
|
||||
? `${base}${EXEC_APPROVAL_FOLLOWUP_ELEVATED_TOKEN_MARKER}${params.execApprovalFollowupToken}`
|
||||
: base;
|
||||
}
|
||||
|
||||
export function parseExecApprovalFollowupApprovalId(idempotencyKey: string): string | undefined {
|
||||
const normalized = normalizeOptionalString(idempotencyKey);
|
||||
if (!normalized?.startsWith(EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_PREFIX)) {
|
||||
function parseExecApprovalFollowupToken(idempotencyKey: string): string | undefined {
|
||||
if (!idempotencyKey.startsWith(EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_PREFIX)) {
|
||||
return undefined;
|
||||
}
|
||||
const body = normalized.slice(EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_PREFIX.length);
|
||||
const nonceMarker = body.lastIndexOf(EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_NONCE_MARKER);
|
||||
return normalizeOptionalString(nonceMarker >= 0 ? body.slice(0, nonceMarker) : body);
|
||||
const tokenMarker = idempotencyKey.lastIndexOf(EXEC_APPROVAL_FOLLOWUP_ELEVATED_TOKEN_MARKER);
|
||||
if (tokenMarker < EXEC_APPROVAL_FOLLOWUP_IDEMPOTENCY_PREFIX.length) {
|
||||
return undefined;
|
||||
}
|
||||
return normalizeOptionalString(
|
||||
idempotencyKey.slice(tokenMarker + EXEC_APPROVAL_FOLLOWUP_ELEVATED_TOKEN_MARKER.length),
|
||||
);
|
||||
}
|
||||
|
||||
export function registerExecApprovalFollowupRuntimeHandoff(params: {
|
||||
approvalId: string;
|
||||
export function registerExecApprovalFollowupElevatedDefaults(params: {
|
||||
sessionKey: string;
|
||||
bashElevated?: ExecElevatedDefaults;
|
||||
nowMs?: number;
|
||||
}): ExecApprovalFollowupRuntimeHandoffRegistration | undefined {
|
||||
const approvalId = normalizeOptionalString(params.approvalId);
|
||||
}): string | undefined {
|
||||
const sessionKey = normalizeOptionalString(params.sessionKey);
|
||||
if (!approvalId || !sessionKey || !params.bashElevated) {
|
||||
if (!params.bashElevated || !sessionKey) {
|
||||
return undefined;
|
||||
}
|
||||
const nowMs = params.nowMs ?? Date.now();
|
||||
pruneExpiredExecApprovalFollowupRuntimeHandoffs(nowMs);
|
||||
const handoffId = randomUUID();
|
||||
const idempotencyKey = buildExecApprovalFollowupIdempotencyKey({
|
||||
approvalId,
|
||||
nonce: randomUUID(),
|
||||
});
|
||||
execApprovalFollowupRuntimeHandoffs.set(handoffId, {
|
||||
kind: "exec-approval-followup",
|
||||
approvalId,
|
||||
pruneExpiredExecApprovalFollowupElevatedDefaults(nowMs);
|
||||
const token = randomUUID();
|
||||
execApprovalFollowupElevatedDefaults.set(token, {
|
||||
sessionKey,
|
||||
idempotencyKey,
|
||||
bashElevated: cloneExecElevatedDefaults(params.bashElevated),
|
||||
expiresAtMs: nowMs + EXEC_APPROVAL_FOLLOWUP_RUNTIME_HANDOFF_TTL_MS,
|
||||
expiresAtMs: nowMs + EXEC_APPROVAL_FOLLOWUP_ELEVATED_TTL_MS,
|
||||
});
|
||||
return { handoffId, idempotencyKey };
|
||||
return token;
|
||||
}
|
||||
|
||||
export function consumeExecApprovalFollowupRuntimeHandoff(params: {
|
||||
handoffId?: string;
|
||||
approvalId?: string;
|
||||
idempotencyKey?: string;
|
||||
export function consumeExecApprovalFollowupElevatedDefaults(params: {
|
||||
token?: string;
|
||||
sessionKey?: string;
|
||||
nowMs?: number;
|
||||
}): ExecApprovalFollowupRuntimeHandoff | undefined {
|
||||
const handoffId = normalizeOptionalString(params.handoffId);
|
||||
const approvalId = normalizeOptionalString(params.approvalId);
|
||||
const idempotencyKey = normalizeOptionalString(params.idempotencyKey);
|
||||
if (!handoffId || !approvalId || !idempotencyKey) {
|
||||
}): ExecElevatedDefaults | undefined {
|
||||
const token = normalizeOptionalString(params.token);
|
||||
if (!token) {
|
||||
return undefined;
|
||||
}
|
||||
const nowMs = params.nowMs ?? Date.now();
|
||||
pruneExpiredExecApprovalFollowupRuntimeHandoffs(nowMs);
|
||||
const entry = execApprovalFollowupRuntimeHandoffs.get(handoffId);
|
||||
pruneExpiredExecApprovalFollowupElevatedDefaults(nowMs);
|
||||
const entry = execApprovalFollowupElevatedDefaults.get(token);
|
||||
if (!entry) {
|
||||
return undefined;
|
||||
}
|
||||
if (entry.expiresAtMs <= nowMs) {
|
||||
execApprovalFollowupRuntimeHandoffs.delete(handoffId);
|
||||
execApprovalFollowupElevatedDefaults.delete(token);
|
||||
return undefined;
|
||||
}
|
||||
const sessionKey = normalizeOptionalString(params.sessionKey);
|
||||
if (
|
||||
entry.approvalId !== approvalId ||
|
||||
entry.idempotencyKey !== idempotencyKey ||
|
||||
entry.sessionKey !== sessionKey
|
||||
) {
|
||||
if (entry.sessionKey !== sessionKey) {
|
||||
return undefined;
|
||||
}
|
||||
execApprovalFollowupRuntimeHandoffs.delete(handoffId);
|
||||
return cloneExecApprovalFollowupRuntimeHandoff(entry);
|
||||
execApprovalFollowupElevatedDefaults.delete(token);
|
||||
return cloneExecElevatedDefaults(entry.bashElevated);
|
||||
}
|
||||
|
||||
export function resetExecApprovalFollowupRuntimeHandoffsForTests(): void {
|
||||
execApprovalFollowupRuntimeHandoffs.clear();
|
||||
export function consumeExecApprovalFollowupElevatedDefaultsFromIdempotencyKey(params: {
|
||||
idempotencyKey: string;
|
||||
sessionKey?: string;
|
||||
nowMs?: number;
|
||||
}): ExecElevatedDefaults | undefined {
|
||||
return consumeExecApprovalFollowupElevatedDefaults({
|
||||
token: parseExecApprovalFollowupToken(params.idempotencyKey),
|
||||
sessionKey: params.sessionKey,
|
||||
nowMs: params.nowMs,
|
||||
});
|
||||
}
|
||||
|
||||
export function resetExecApprovalFollowupElevatedDefaultsForTests(): void {
|
||||
execApprovalFollowupElevatedDefaults.clear();
|
||||
}
|
||||
|
||||
@@ -283,14 +283,13 @@ describe("exec approval followup", () => {
|
||||
expect(sendMessage).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it("carries the runtime handoff separately from idempotency without exposing elevated defaults", async () => {
|
||||
it("carries the elevated followup token through idempotency without exposing elevated defaults", async () => {
|
||||
await sendExecApprovalFollowup({
|
||||
approvalId: "req-elevated-75832",
|
||||
sessionKey: "agent:main:telegram:direct:123",
|
||||
turnSourceChannel: "telegram",
|
||||
resultText: "Exec finished (gateway id=req-elevated-75832, code 0)\nok",
|
||||
internalRuntimeHandoffId: "handoff-75832",
|
||||
idempotencyKey: "exec-approval-followup:req-elevated-75832:nonce:nonce-75832",
|
||||
execApprovalFollowupToken: "token-75832",
|
||||
});
|
||||
|
||||
expect(callGatewayTool).toHaveBeenCalledWith(
|
||||
@@ -299,8 +298,7 @@ describe("exec approval followup", () => {
|
||||
expect.objectContaining({
|
||||
sessionKey: "agent:main:telegram:direct:123",
|
||||
channel: "telegram",
|
||||
idempotencyKey: "exec-approval-followup:req-elevated-75832:nonce:nonce-75832",
|
||||
internalRuntimeHandoffId: "handoff-75832",
|
||||
idempotencyKey: "exec-approval-followup:req-elevated-75832:elevated:token-75832",
|
||||
}),
|
||||
{ expectFinal: true },
|
||||
);
|
||||
|
||||
@@ -24,8 +24,7 @@ type ExecApprovalFollowupParams = {
|
||||
turnSourceThreadId?: string | number;
|
||||
resultText: string;
|
||||
direct?: boolean;
|
||||
internalRuntimeHandoffId?: string;
|
||||
idempotencyKey?: string;
|
||||
execApprovalFollowupToken?: string;
|
||||
};
|
||||
|
||||
function buildExecDeniedFollowupPrompt(resultText: string): string {
|
||||
@@ -152,8 +151,7 @@ function buildAgentFollowupArgs(params: {
|
||||
turnSourceTo?: string;
|
||||
turnSourceAccountId?: string;
|
||||
turnSourceThreadId?: string | number;
|
||||
internalRuntimeHandoffId?: string;
|
||||
idempotencyKey?: string;
|
||||
execApprovalFollowupToken?: string;
|
||||
}) {
|
||||
const { deliveryTarget, sessionOnlyOriginChannel } = params;
|
||||
// When the followup run has no deliverable route and no gateway-internal channel,
|
||||
@@ -181,14 +179,10 @@ function buildAgentFollowupArgs(params: {
|
||||
: sessionOnlyOriginChannel
|
||||
? params.turnSourceThreadId
|
||||
: undefined,
|
||||
idempotencyKey:
|
||||
params.idempotencyKey ??
|
||||
buildExecApprovalFollowupIdempotencyKey({
|
||||
approvalId: params.approvalId,
|
||||
}),
|
||||
...(params.internalRuntimeHandoffId
|
||||
? { internalRuntimeHandoffId: params.internalRuntimeHandoffId }
|
||||
: {}),
|
||||
idempotencyKey: buildExecApprovalFollowupIdempotencyKey({
|
||||
approvalId: params.approvalId,
|
||||
execApprovalFollowupToken: params.execApprovalFollowupToken,
|
||||
}),
|
||||
};
|
||||
}
|
||||
|
||||
@@ -262,8 +256,7 @@ export async function sendExecApprovalFollowup(
|
||||
turnSourceTo: params.turnSourceTo,
|
||||
turnSourceAccountId: params.turnSourceAccountId,
|
||||
turnSourceThreadId: params.turnSourceThreadId,
|
||||
internalRuntimeHandoffId: params.internalRuntimeHandoffId,
|
||||
idempotencyKey: params.idempotencyKey,
|
||||
execApprovalFollowupToken: params.execApprovalFollowupToken,
|
||||
}),
|
||||
{ expectFinal: true },
|
||||
);
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
import { beforeEach, describe, expect, it, vi } from "vitest";
|
||||
import {
|
||||
consumeExecApprovalFollowupRuntimeHandoff,
|
||||
resetExecApprovalFollowupRuntimeHandoffsForTests,
|
||||
consumeExecApprovalFollowupElevatedDefaults,
|
||||
resetExecApprovalFollowupElevatedDefaultsForTests,
|
||||
} from "./bash-tools.exec-approval-followup-state.js";
|
||||
import {
|
||||
buildExecApprovalPendingToolResult,
|
||||
@@ -63,7 +63,7 @@ describe("sendExecApprovalFollowupResult", () => {
|
||||
allowlist: [],
|
||||
file: { version: 1, agents: {} },
|
||||
});
|
||||
resetExecApprovalFollowupRuntimeHandoffsForTests();
|
||||
resetExecApprovalFollowupElevatedDefaultsForTests();
|
||||
});
|
||||
|
||||
it("logs repeated followup dispatch failures once per approval id and error message", async () => {
|
||||
@@ -132,65 +132,22 @@ describe("sendExecApprovalFollowupResult", () => {
|
||||
);
|
||||
|
||||
const call = sendExecApprovalFollowup.mock.calls[0]?.[0] as
|
||||
| {
|
||||
internalRuntimeHandoffId?: string;
|
||||
idempotencyKey?: string;
|
||||
execApprovalFollowupToken?: string;
|
||||
bashElevated?: unknown;
|
||||
}
|
||||
| { execApprovalFollowupToken?: string; bashElevated?: unknown }
|
||||
| undefined;
|
||||
expect(call?.internalRuntimeHandoffId).toEqual(expect.any(String));
|
||||
expect(call?.idempotencyKey).toMatch(/^exec-approval-followup:approval-elevated-75832:nonce:/);
|
||||
expect(call?.idempotencyKey).not.toContain(call?.internalRuntimeHandoffId ?? "");
|
||||
expect(call?.execApprovalFollowupToken).toEqual(expect.any(String));
|
||||
expect(call).not.toHaveProperty("bashElevated");
|
||||
expect(call).not.toHaveProperty("execApprovalFollowupToken");
|
||||
expect(
|
||||
consumeExecApprovalFollowupRuntimeHandoff({
|
||||
handoffId: call?.internalRuntimeHandoffId ?? "",
|
||||
approvalId: "approval-elevated-75832",
|
||||
idempotencyKey: call?.idempotencyKey ?? "",
|
||||
consumeExecApprovalFollowupElevatedDefaults({
|
||||
token: call?.execApprovalFollowupToken ?? "",
|
||||
sessionKey: "agent:main:telegram:direct:wrong",
|
||||
}),
|
||||
).toBeUndefined();
|
||||
expect(
|
||||
consumeExecApprovalFollowupRuntimeHandoff({
|
||||
handoffId: call?.internalRuntimeHandoffId ?? "",
|
||||
approvalId: "approval-elevated-75832",
|
||||
idempotencyKey: call?.idempotencyKey ?? "",
|
||||
consumeExecApprovalFollowupElevatedDefaults({
|
||||
token: call?.execApprovalFollowupToken ?? "",
|
||||
sessionKey: "agent:main:telegram:direct:123",
|
||||
}),
|
||||
).toEqual({
|
||||
kind: "exec-approval-followup",
|
||||
approvalId: "approval-elevated-75832",
|
||||
sessionKey: "agent:main:telegram:direct:123",
|
||||
idempotencyKey: call?.idempotencyKey,
|
||||
bashElevated,
|
||||
});
|
||||
});
|
||||
|
||||
it("keeps non-elevated agent followups on the deterministic idempotency path", async () => {
|
||||
sendExecApprovalFollowup.mockResolvedValue(true);
|
||||
|
||||
await sendExecApprovalFollowupResult(
|
||||
{
|
||||
approvalId: "approval-normal-75832",
|
||||
sessionKey: "agent:main:telegram:direct:123",
|
||||
turnSourceChannel: "telegram",
|
||||
},
|
||||
"Exec finished",
|
||||
{ sendExecApprovalFollowup, logWarn },
|
||||
);
|
||||
|
||||
const call = sendExecApprovalFollowup.mock.calls[0]?.[0] as
|
||||
| {
|
||||
internalRuntimeHandoffId?: string;
|
||||
idempotencyKey?: string;
|
||||
bashElevated?: unknown;
|
||||
}
|
||||
| undefined;
|
||||
expect(call).not.toHaveProperty("internalRuntimeHandoffId");
|
||||
expect(call).not.toHaveProperty("idempotencyKey");
|
||||
expect(call).not.toHaveProperty("bashElevated");
|
||||
).toEqual(bashElevated);
|
||||
});
|
||||
});
|
||||
|
||||
|
||||
@@ -16,7 +16,7 @@ import {
|
||||
type ExecSecurity,
|
||||
} from "../infra/exec-approvals.js";
|
||||
import { logWarn } from "../logger.js";
|
||||
import { registerExecApprovalFollowupRuntimeHandoff } from "./bash-tools.exec-approval-followup-state.js";
|
||||
import { registerExecApprovalFollowupElevatedDefaults } from "./bash-tools.exec-approval-followup-state.js";
|
||||
import { sendExecApprovalFollowup } from "./bash-tools.exec-approval-followup.js";
|
||||
import {
|
||||
type ExecApprovalRegistration,
|
||||
@@ -411,11 +411,10 @@ export async function sendExecApprovalFollowupResult(
|
||||
): Promise<void> {
|
||||
const send = deps.sendExecApprovalFollowup ?? sendExecApprovalFollowup;
|
||||
const warn = deps.logWarn ?? logWarn;
|
||||
const runtimeHandoff =
|
||||
const execApprovalFollowupToken =
|
||||
target.direct === true || !target.sessionKey
|
||||
? undefined
|
||||
: registerExecApprovalFollowupRuntimeHandoff({
|
||||
approvalId: target.approvalId,
|
||||
: registerExecApprovalFollowupElevatedDefaults({
|
||||
sessionKey: target.sessionKey,
|
||||
bashElevated: target.bashElevated,
|
||||
});
|
||||
@@ -428,12 +427,7 @@ export async function sendExecApprovalFollowupResult(
|
||||
turnSourceThreadId: target.turnSourceThreadId,
|
||||
resultText,
|
||||
direct: target.direct,
|
||||
...(runtimeHandoff
|
||||
? {
|
||||
internalRuntimeHandoffId: runtimeHandoff.handoffId,
|
||||
idempotencyKey: runtimeHandoff.idempotencyKey,
|
||||
}
|
||||
: {}),
|
||||
execApprovalFollowupToken,
|
||||
}).catch((error) => {
|
||||
const message = formatErrorMessage(error);
|
||||
const key = `${target.approvalId}:${message}`;
|
||||
|
||||
@@ -1610,7 +1610,7 @@ export function createExecTool(
|
||||
type: "text",
|
||||
text: `${getWarningText()}Command still running (session ${run.session.id}, pid ${
|
||||
run.session.pid ?? "n/a"
|
||||
}). Use process (list/poll/log/write/send-keys/submit/paste/kill/clear/remove) for follow-up.`,
|
||||
}). Use process (list/poll/log/write/kill/clear/remove) for follow-up.`,
|
||||
},
|
||||
],
|
||||
details: {
|
||||
|
||||
@@ -7,9 +7,6 @@ export type WritableStdin = {
|
||||
write: (data: string, cb?: (err?: Error | null) => void) => void;
|
||||
end: () => void;
|
||||
destroyed?: boolean;
|
||||
writable?: boolean;
|
||||
writableEnded?: boolean;
|
||||
writableFinished?: boolean;
|
||||
};
|
||||
|
||||
function failText(text: string): AgentToolResult<unknown> {
|
||||
|
||||
@@ -1,225 +0,0 @@
|
||||
import { afterEach, describe, expect, it, vi } from "vitest";
|
||||
import {
|
||||
addSession,
|
||||
appendOutput,
|
||||
markExited,
|
||||
resetProcessRegistryForTests,
|
||||
} from "./bash-process-registry.js";
|
||||
import { createProcessSessionFixture } from "./bash-process-registry.test-helpers.js";
|
||||
import { createProcessTool } from "./bash-tools.process.js";
|
||||
|
||||
type ProcessTool = ReturnType<typeof createProcessTool>;
|
||||
type ProcessToolResult = Awaited<ReturnType<ProcessTool["execute"]>>;
|
||||
|
||||
afterEach(() => {
|
||||
resetProcessRegistryForTests();
|
||||
vi.useRealTimers();
|
||||
});
|
||||
|
||||
async function runProcessAction(
|
||||
processTool: ProcessTool,
|
||||
args: Record<string, unknown>,
|
||||
): Promise<ProcessToolResult> {
|
||||
return processTool.execute("toolcall", args as Parameters<ProcessTool["execute"]>[1], undefined);
|
||||
}
|
||||
|
||||
function textOf(result: ProcessToolResult): string {
|
||||
const item = result.content[0];
|
||||
return item?.type === "text" ? item.text : "";
|
||||
}
|
||||
|
||||
function installWritableStdin(
|
||||
session: ReturnType<typeof createProcessSessionFixture>,
|
||||
state?: { writableEnded?: boolean; writableFinished?: boolean; destroyed?: boolean },
|
||||
) {
|
||||
session.stdin = {
|
||||
write: vi.fn((_data: string, cb?: (err?: Error | null) => void) => cb?.(null)),
|
||||
end: vi.fn(),
|
||||
destroyed: state?.destroyed ?? false,
|
||||
writableEnded: state?.writableEnded,
|
||||
writableFinished: state?.writableFinished,
|
||||
} as NonNullable<typeof session.stdin> & {
|
||||
writableEnded?: boolean;
|
||||
writableFinished?: boolean;
|
||||
};
|
||||
}
|
||||
|
||||
describe("process input-wait hints", () => {
|
||||
it("adds output and input-wait metadata to log for an idle writable session", async () => {
|
||||
vi.useFakeTimers();
|
||||
vi.setSystemTime(new Date("2026-01-01T00:00:20.000Z"));
|
||||
const processTool = createProcessTool();
|
||||
const session = createProcessSessionFixture({
|
||||
id: "sess-log-hint",
|
||||
command: "node cli.js",
|
||||
backgrounded: true,
|
||||
startedAt: Date.now() - 20_000,
|
||||
});
|
||||
installWritableStdin(session);
|
||||
appendOutput(session, "stdout", "Name? ");
|
||||
addSession(session);
|
||||
|
||||
const result = await runProcessAction(processTool, {
|
||||
action: "log",
|
||||
sessionId: "sess-log-hint",
|
||||
});
|
||||
|
||||
const text = textOf(result);
|
||||
expect(text).toContain("Name? ");
|
||||
expect(text).toContain("No new output for 20s");
|
||||
expect(text).toContain("Use process write, send-keys, submit, or paste to provide input.");
|
||||
expect(result.details).toMatchObject({
|
||||
status: "running",
|
||||
sessionId: "sess-log-hint",
|
||||
stdinWritable: true,
|
||||
waitingForInput: true,
|
||||
idleMs: 20_000,
|
||||
lastOutputAt: Date.now() - 20_000,
|
||||
});
|
||||
});
|
||||
|
||||
it("adds input-wait hints to poll when no new output arrives", async () => {
|
||||
vi.useFakeTimers();
|
||||
vi.setSystemTime(new Date("2026-01-01T00:00:16.000Z"));
|
||||
const processTool = createProcessTool();
|
||||
const session = createProcessSessionFixture({
|
||||
id: "sess-poll",
|
||||
command: "python prompt.py",
|
||||
backgrounded: true,
|
||||
startedAt: Date.now() - 16_000,
|
||||
});
|
||||
installWritableStdin(session);
|
||||
addSession(session);
|
||||
|
||||
const result = await runProcessAction(processTool, {
|
||||
action: "poll",
|
||||
sessionId: "sess-poll",
|
||||
});
|
||||
|
||||
expect(textOf(result)).toContain("(no new output)");
|
||||
expect(textOf(result)).toContain("may be waiting for input");
|
||||
expect(result.details).toMatchObject({
|
||||
status: "running",
|
||||
sessionId: "sess-poll",
|
||||
stdinWritable: true,
|
||||
waitingForInput: true,
|
||||
idleMs: 16_000,
|
||||
lastOutputAt: Date.now() - 16_000,
|
||||
});
|
||||
});
|
||||
|
||||
it("marks idle writable sessions in process list", async () => {
|
||||
vi.useFakeTimers();
|
||||
vi.setSystemTime(new Date("2026-01-01T00:00:30.000Z"));
|
||||
const processTool = createProcessTool();
|
||||
const session = createProcessSessionFixture({
|
||||
id: "sess-list",
|
||||
command: "npm run interactive",
|
||||
backgrounded: true,
|
||||
startedAt: Date.now() - 30_000,
|
||||
});
|
||||
installWritableStdin(session);
|
||||
addSession(session);
|
||||
|
||||
const result = await runProcessAction(processTool, { action: "list" });
|
||||
|
||||
expect(textOf(result)).toContain("sess-list");
|
||||
expect(textOf(result)).toContain("[input-wait]");
|
||||
const sessions = (result.details as { sessions?: Array<Record<string, unknown>> }).sessions;
|
||||
expect(sessions?.[0]).toMatchObject({
|
||||
sessionId: "sess-list",
|
||||
stdinWritable: true,
|
||||
waitingForInput: true,
|
||||
idleMs: 30_000,
|
||||
});
|
||||
});
|
||||
|
||||
it("adds input-wait metadata and hint text to log", async () => {
|
||||
vi.useFakeTimers();
|
||||
vi.setSystemTime(new Date("2026-01-01T00:00:25.000Z"));
|
||||
const processTool = createProcessTool();
|
||||
const session = createProcessSessionFixture({
|
||||
id: "sess-log",
|
||||
command: "node prompt.js",
|
||||
backgrounded: true,
|
||||
startedAt: Date.now() - 25_000,
|
||||
});
|
||||
installWritableStdin(session);
|
||||
appendOutput(session, "stdout", "Password: ");
|
||||
addSession(session);
|
||||
|
||||
const result = await runProcessAction(processTool, {
|
||||
action: "log",
|
||||
sessionId: "sess-log",
|
||||
});
|
||||
|
||||
expect(textOf(result)).toContain("Password: ");
|
||||
expect(textOf(result)).toContain("No new output for 25s");
|
||||
expect(textOf(result)).toContain("Use process write, send-keys, submit, or paste");
|
||||
expect(result.details).toMatchObject({
|
||||
status: "running",
|
||||
sessionId: "sess-log",
|
||||
stdinWritable: true,
|
||||
waitingForInput: true,
|
||||
idleMs: 25_000,
|
||||
});
|
||||
});
|
||||
|
||||
it("does not treat ended stdin as writable input-wait state", async () => {
|
||||
vi.useFakeTimers();
|
||||
vi.setSystemTime(new Date("2026-01-01T00:01:00.000Z"));
|
||||
const processTool = createProcessTool();
|
||||
const session = createProcessSessionFixture({
|
||||
id: "sess-ended",
|
||||
command: "node closed-stdin.js",
|
||||
backgrounded: true,
|
||||
startedAt: Date.now() - 60_000,
|
||||
});
|
||||
installWritableStdin(session, { writableEnded: true });
|
||||
addSession(session);
|
||||
|
||||
const log = await runProcessAction(processTool, {
|
||||
action: "log",
|
||||
sessionId: "sess-ended",
|
||||
});
|
||||
expect(textOf(log)).not.toContain("provide input");
|
||||
expect(log.details).toMatchObject({
|
||||
status: "running",
|
||||
stdinWritable: false,
|
||||
waitingForInput: false,
|
||||
});
|
||||
|
||||
const write = await runProcessAction(processTool, {
|
||||
action: "write",
|
||||
sessionId: "sess-ended",
|
||||
data: "answer\n",
|
||||
});
|
||||
expect(textOf(write)).toContain("stdin is not writable");
|
||||
expect(write.details).toMatchObject({ status: "failed" });
|
||||
});
|
||||
|
||||
it("can read finished session logs without exposing input controls", async () => {
|
||||
const processTool = createProcessTool();
|
||||
const session = createProcessSessionFixture({
|
||||
id: "sess-finished",
|
||||
command: "echo done",
|
||||
backgrounded: true,
|
||||
});
|
||||
appendOutput(session, "stdout", "done\n");
|
||||
addSession(session);
|
||||
markExited(session, 0, null, "completed");
|
||||
|
||||
const result = await runProcessAction(processTool, {
|
||||
action: "log",
|
||||
sessionId: "sess-finished",
|
||||
});
|
||||
|
||||
expect(textOf(result)).toContain("done");
|
||||
expect(textOf(result)).not.toContain("provide input");
|
||||
expect(result.details).toMatchObject({
|
||||
status: "completed",
|
||||
sessionId: "sess-finished",
|
||||
exitCode: 0,
|
||||
});
|
||||
});
|
||||
});
|
||||
@@ -17,14 +17,7 @@ import {
|
||||
import { describeProcessTool } from "./bash-tools.descriptions.js";
|
||||
import { handleProcessSendKeys, type WritableStdin } from "./bash-tools.process-send-keys.js";
|
||||
import { processSchema } from "./bash-tools.schemas.js";
|
||||
import {
|
||||
clampWithDefault,
|
||||
deriveSessionName,
|
||||
pad,
|
||||
readEnvInt,
|
||||
sliceLogLines,
|
||||
truncateMiddle,
|
||||
} from "./bash-tools.shared.js";
|
||||
import { deriveSessionName, pad, sliceLogLines, truncateMiddle } from "./bash-tools.shared.js";
|
||||
import { recordCommandPoll, resetCommandPollCount } from "./command-poll-backoff.js";
|
||||
import { encodePaste } from "./pty-keys.js";
|
||||
import { PROCESS_TOOL_DISPLAY_SUMMARY } from "./tool-description-presets.js";
|
||||
@@ -33,14 +26,10 @@ import type { AgentToolWithMeta } from "./tools/common.js";
|
||||
export type ProcessToolDefaults = {
|
||||
cleanupMs?: number;
|
||||
hasCronTool?: boolean;
|
||||
inputWaitIdleMs?: number;
|
||||
scopeKey?: string;
|
||||
};
|
||||
|
||||
const DEFAULT_LOG_TAIL_LINES = 200;
|
||||
const DEFAULT_INPUT_WAIT_IDLE_MS = 15_000;
|
||||
const MIN_INPUT_WAIT_IDLE_MS = 1_000;
|
||||
const MAX_INPUT_WAIT_IDLE_MS = 10 * 60 * 1000;
|
||||
|
||||
function resolveLogSliceWindow(offset?: number, limit?: number) {
|
||||
const usingDefaultTail = offset === undefined && limit === undefined;
|
||||
@@ -62,36 +51,6 @@ function defaultTailNote(totalLines: number, usingDefaultTail: boolean) {
|
||||
|
||||
const MAX_POLL_WAIT_MS = 30_000;
|
||||
|
||||
type RunningSessionRuntime = {
|
||||
stdinWritable: boolean;
|
||||
waitingForInput: boolean;
|
||||
idleMs: number;
|
||||
lastOutputAt: number;
|
||||
};
|
||||
|
||||
function resolveSessionStdin(session: ProcessSession): WritableStdin | undefined {
|
||||
return (session.stdin ?? session.child?.stdin) as WritableStdin | undefined;
|
||||
}
|
||||
|
||||
function isWritableStdin(stdin: WritableStdin | undefined): stdin is WritableStdin {
|
||||
if (!stdin || stdin.destroyed) {
|
||||
return false;
|
||||
}
|
||||
if (stdin.writable === false || stdin.writableEnded === true || stdin.writableFinished === true) {
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
function runningSessionInputDetails(runtime: RunningSessionRuntime) {
|
||||
return {
|
||||
stdinWritable: runtime.stdinWritable,
|
||||
waitingForInput: runtime.waitingForInput,
|
||||
idleMs: runtime.idleMs,
|
||||
lastOutputAt: runtime.lastOutputAt,
|
||||
};
|
||||
}
|
||||
|
||||
function resolvePollWaitMs(value: unknown) {
|
||||
if (typeof value === "number" && Number.isFinite(value)) {
|
||||
return Math.max(0, Math.min(MAX_POLL_WAIT_MS, Math.floor(value)));
|
||||
@@ -181,36 +140,9 @@ export function createProcessTool(
|
||||
}
|
||||
const scopeKey = defaults?.scopeKey;
|
||||
const supervisor = getProcessSupervisor();
|
||||
const inputWaitIdleMs = clampWithDefault(
|
||||
defaults?.inputWaitIdleMs ?? readEnvInt("OPENCLAW_PROCESS_INPUT_WAIT_IDLE_MS"),
|
||||
DEFAULT_INPUT_WAIT_IDLE_MS,
|
||||
MIN_INPUT_WAIT_IDLE_MS,
|
||||
MAX_INPUT_WAIT_IDLE_MS,
|
||||
);
|
||||
const isInScope = (session?: { scopeKey?: string } | null) =>
|
||||
!scopeKey || session?.scopeKey === scopeKey;
|
||||
|
||||
const describeRunningSession = (session: ProcessSession): RunningSessionRuntime => {
|
||||
const record = supervisor.getRecord(session.id);
|
||||
const lastOutputAt = record?.lastOutputAtMs ?? session.startedAt;
|
||||
const idleMs = Math.max(0, Date.now() - lastOutputAt);
|
||||
const stdinWritable = isWritableStdin(resolveSessionStdin(session));
|
||||
return {
|
||||
stdinWritable,
|
||||
waitingForInput: stdinWritable && idleMs >= inputWaitIdleMs,
|
||||
idleMs,
|
||||
lastOutputAt,
|
||||
};
|
||||
};
|
||||
|
||||
const buildInputWaitHint = (runtime: RunningSessionRuntime | undefined) => {
|
||||
if (!runtime?.waitingForInput) {
|
||||
return "";
|
||||
}
|
||||
const idle = formatDurationCompact(runtime.idleMs) ?? `${runtime.idleMs}ms`;
|
||||
return `\n\nNo new output for ${idle}; this session may be waiting for input. Use process write, send-keys, submit, or paste to provide input.`;
|
||||
};
|
||||
|
||||
const cancelManagedSession = (sessionId: string) => {
|
||||
const record = supervisor.getRecord(sessionId);
|
||||
if (!record || record.state === "exited") {
|
||||
@@ -264,25 +196,18 @@ export function createProcessTool(
|
||||
if (params.action === "list") {
|
||||
const running = listRunningSessions()
|
||||
.filter((s) => isInScope(s))
|
||||
.map((s) => {
|
||||
const runtime = describeRunningSession(s);
|
||||
return {
|
||||
sessionId: s.id,
|
||||
status: "running",
|
||||
pid: s.pid ?? undefined,
|
||||
startedAt: s.startedAt,
|
||||
runtimeMs: Date.now() - s.startedAt,
|
||||
cwd: s.cwd,
|
||||
command: s.command,
|
||||
name: deriveSessionName(s.command),
|
||||
tail: s.tail,
|
||||
truncated: s.truncated,
|
||||
stdinWritable: runtime.stdinWritable,
|
||||
waitingForInput: runtime.waitingForInput,
|
||||
idleMs: runtime.idleMs,
|
||||
lastOutputAt: runtime.lastOutputAt,
|
||||
};
|
||||
});
|
||||
.map((s) => ({
|
||||
sessionId: s.id,
|
||||
status: "running",
|
||||
pid: s.pid ?? undefined,
|
||||
startedAt: s.startedAt,
|
||||
runtimeMs: Date.now() - s.startedAt,
|
||||
cwd: s.cwd,
|
||||
command: s.command,
|
||||
name: deriveSessionName(s.command),
|
||||
tail: s.tail,
|
||||
truncated: s.truncated,
|
||||
}));
|
||||
const finished = listFinishedSessions()
|
||||
.filter((s) => isInScope(s))
|
||||
.map((s) => ({
|
||||
@@ -303,10 +228,7 @@ export function createProcessTool(
|
||||
.toSorted((a, b) => b.startedAt - a.startedAt)
|
||||
.map((s) => {
|
||||
const label = s.name ? truncateMiddle(s.name, 80) : truncateMiddle(s.command, 120);
|
||||
const marker = "waitingForInput" in s && s.waitingForInput ? " [input-wait]" : "";
|
||||
return `${s.sessionId} ${pad(s.status, 9)} ${
|
||||
formatDurationCompact(s.runtimeMs) ?? "n/a"
|
||||
}${marker} :: ${label}`;
|
||||
return `${s.sessionId} ${pad(s.status, 9)} ${formatDurationCompact(s.runtimeMs) ?? "n/a"} :: ${label}`;
|
||||
});
|
||||
return {
|
||||
content: [
|
||||
@@ -349,14 +271,14 @@ export function createProcessTool(
|
||||
result: failedResult(`Session ${params.sessionId} is not backgrounded.`),
|
||||
};
|
||||
}
|
||||
const stdin = resolveSessionStdin(scopedSession);
|
||||
if (!isWritableStdin(stdin)) {
|
||||
const stdin = scopedSession.stdin ?? scopedSession.child?.stdin;
|
||||
if (!stdin || stdin.destroyed) {
|
||||
return {
|
||||
ok: false as const,
|
||||
result: failedResult(`Session ${params.sessionId} stdin is not writable.`),
|
||||
};
|
||||
}
|
||||
return { ok: true as const, session: scopedSession, stdin };
|
||||
return { ok: true as const, session: scopedSession, stdin: stdin as WritableStdin };
|
||||
};
|
||||
|
||||
const writeToStdin = async (stdin: WritableStdin, data: string) => {
|
||||
@@ -452,7 +374,6 @@ export function createProcessTool(
|
||||
if (exited) {
|
||||
resetPollRetrySuggestion(params.sessionId);
|
||||
}
|
||||
const runtime = exited ? undefined : describeRunningSession(scopedSession);
|
||||
return {
|
||||
content: [
|
||||
{
|
||||
@@ -463,7 +384,7 @@ export function createProcessTool(
|
||||
? `\n\nProcess exited with ${
|
||||
exitSignal ? `signal ${exitSignal}` : `code ${exitCode}`
|
||||
}.`
|
||||
: buildInputWaitHint(runtime) || "\n\nProcess still running."),
|
||||
: "\n\nProcess still running."),
|
||||
},
|
||||
],
|
||||
details: {
|
||||
@@ -472,7 +393,6 @@ export function createProcessTool(
|
||||
exitCode: exited ? exitCode : undefined,
|
||||
aggregated: scopedSession.aggregated,
|
||||
name: deriveSessionName(scopedSession.command),
|
||||
...(runtime ? runningSessionInputDetails(runtime) : {}),
|
||||
...(typeof retryInMs === "number" ? { retryInMs } : {}),
|
||||
},
|
||||
};
|
||||
@@ -497,16 +417,9 @@ export function createProcessTool(
|
||||
window.effectiveOffset,
|
||||
window.effectiveLimit,
|
||||
);
|
||||
const runtime = describeRunningSession(scopedSession);
|
||||
const logDefaultTailNote = defaultTailNote(totalLines, window.usingDefaultTail);
|
||||
return {
|
||||
content: [
|
||||
{
|
||||
type: "text",
|
||||
text:
|
||||
(slice || "(no output yet)") + logDefaultTailNote + buildInputWaitHint(runtime),
|
||||
},
|
||||
],
|
||||
content: [{ type: "text", text: (slice || "(no output yet)") + logDefaultTailNote }],
|
||||
details: {
|
||||
status: scopedSession.exited ? "completed" : "running",
|
||||
sessionId: params.sessionId,
|
||||
@@ -515,7 +428,6 @@ export function createProcessTool(
|
||||
totalChars,
|
||||
truncated: scopedSession.truncated,
|
||||
name: deriveSessionName(scopedSession.command),
|
||||
...runningSessionInputDetails(runtime),
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
@@ -50,9 +50,7 @@ export const execSchema = Type.Object({
|
||||
});
|
||||
|
||||
export const processSchema = Type.Object({
|
||||
action: Type.String({
|
||||
description: "Process action (list|poll|log|write|send-keys|submit|paste|kill|clear|remove)",
|
||||
}),
|
||||
action: Type.String({ description: "Process action" }),
|
||||
sessionId: Type.Optional(Type.String({ description: "Session id for actions other than list" })),
|
||||
data: Type.Optional(Type.String({ description: "Data to write for write" })),
|
||||
keys: Type.Optional(
|
||||
|
||||
@@ -249,51 +249,6 @@ async function runMathSideQuestionAndCaptureContext() {
|
||||
return context;
|
||||
}
|
||||
|
||||
function expectRecordFields(
|
||||
record: unknown,
|
||||
expected: Record<string, unknown>,
|
||||
): Record<string, unknown> {
|
||||
expect(record).toBeDefined();
|
||||
const actual = record as Record<string, unknown>;
|
||||
for (const [key, value] of Object.entries(expected)) {
|
||||
expect(actual[key]).toEqual(value);
|
||||
}
|
||||
return actual;
|
||||
}
|
||||
|
||||
function streamContext(callIndex = 0): {
|
||||
messages?: Array<Record<string, unknown>>;
|
||||
systemPrompt?: unknown;
|
||||
} {
|
||||
const call = streamSimpleMock.mock.calls[callIndex];
|
||||
expect(call).toBeDefined();
|
||||
return (call?.[1] ?? {}) as {
|
||||
messages?: Array<Record<string, unknown>>;
|
||||
systemPrompt?: unknown;
|
||||
};
|
||||
}
|
||||
|
||||
function contextMessages(context: unknown): Array<Record<string, unknown>> {
|
||||
const messages = (context as { messages?: Array<Record<string, unknown>> }).messages;
|
||||
expect(messages).toBeDefined();
|
||||
return messages ?? [];
|
||||
}
|
||||
|
||||
function expectTextBlockContains(block: unknown, text: string): void {
|
||||
const record = expectRecordFields(block, { type: "text" });
|
||||
expect(typeof record.text).toBe("string");
|
||||
expect(record.text).toContain(text);
|
||||
}
|
||||
|
||||
function firstTextBlockIncludes(message: Record<string, unknown>, text: string): boolean {
|
||||
if (!Array.isArray(message.content)) {
|
||||
return false;
|
||||
}
|
||||
const [block] = message.content;
|
||||
const blockText = (block as { text?: unknown } | undefined)?.text;
|
||||
return typeof blockText === "string" && blockText.includes(text);
|
||||
}
|
||||
|
||||
function expectNoAssistantMessages(context: unknown) {
|
||||
expect(
|
||||
(context as { messages?: Array<{ role?: string }> }).messages?.filter(
|
||||
@@ -303,24 +258,28 @@ function expectNoAssistantMessages(context: unknown) {
|
||||
}
|
||||
|
||||
function expectSanitizedAssistantContext(context: unknown, text: string) {
|
||||
const messages = contextMessages(context);
|
||||
expect(messages).toHaveLength(3);
|
||||
expectRecordFields(messages[0], { role: "user" });
|
||||
expectRecordFields(messages[1], {
|
||||
role: "assistant",
|
||||
content: [{ type: "text", text }],
|
||||
expect(context).toMatchObject({
|
||||
messages: [
|
||||
expect.objectContaining({ role: "user" }),
|
||||
expect.objectContaining({
|
||||
role: "assistant",
|
||||
content: [{ type: "text", text }],
|
||||
}),
|
||||
expect.objectContaining({ role: "user" }),
|
||||
],
|
||||
});
|
||||
expectRecordFields(messages[2], { role: "user" });
|
||||
}
|
||||
|
||||
function expectSeedOnlyUserContext(context: unknown) {
|
||||
const messages = contextMessages(context);
|
||||
expect(messages).toHaveLength(2);
|
||||
expectRecordFields(messages[0], {
|
||||
role: "user",
|
||||
content: [{ type: "text", text: "seed" }],
|
||||
expect(context).toMatchObject({
|
||||
messages: [
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content: [{ type: "text", text: "seed" }],
|
||||
}),
|
||||
expect.objectContaining({ role: "user" }),
|
||||
],
|
||||
});
|
||||
expectRecordFields(messages[1], { role: "user" });
|
||||
}
|
||||
|
||||
describe("runBtwSideQuestion", () => {
|
||||
@@ -466,9 +425,13 @@ describe("runBtwSideQuestion", () => {
|
||||
const result = await runSideQuestion();
|
||||
|
||||
expect(result).toEqual({ text: "Final answer." });
|
||||
const ensureArgs = ensureOpenClawModelsJsonMock.mock.calls[0];
|
||||
expect(ensureArgs?.[1]).toBe(DEFAULT_AGENT_DIR);
|
||||
expect(ensureArgs?.[2]).toEqual({ workspaceDir: "/tmp/workspace" });
|
||||
expect(ensureOpenClawModelsJsonMock).toHaveBeenCalledWith(
|
||||
expect.any(Object),
|
||||
DEFAULT_AGENT_DIR,
|
||||
{
|
||||
workspaceDir: "/tmp/workspace",
|
||||
},
|
||||
);
|
||||
});
|
||||
|
||||
it("applies provider runtime auth before streaming github-copilot BTW questions", async () => {
|
||||
@@ -497,28 +460,29 @@ describe("runBtwSideQuestion", () => {
|
||||
});
|
||||
|
||||
expect(result).toEqual({ text: "Copilot answer." });
|
||||
const runtimeAuthParams = expectRecordFields(
|
||||
prepareProviderRuntimeAuthMock.mock.calls[0]?.[0],
|
||||
{
|
||||
expect(prepareProviderRuntimeAuthMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
provider: "github-copilot",
|
||||
workspaceDir: "/tmp/workspace",
|
||||
},
|
||||
context: expect.objectContaining({
|
||||
provider: "github-copilot",
|
||||
modelId: "gpt-5.4",
|
||||
workspaceDir: "/tmp/workspace",
|
||||
apiKey: "github-token",
|
||||
authMode: "token",
|
||||
profileId: "profile-1",
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(streamSimpleMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
provider: "github-copilot",
|
||||
id: "gpt-5.4",
|
||||
baseUrl: "https://api.enterprise.githubcopilot.com",
|
||||
}),
|
||||
expect.anything(),
|
||||
expect.objectContaining({ apiKey: "copilot-runtime-token" }),
|
||||
);
|
||||
expectRecordFields(runtimeAuthParams.context, {
|
||||
provider: "github-copilot",
|
||||
modelId: "gpt-5.4",
|
||||
workspaceDir: "/tmp/workspace",
|
||||
apiKey: "github-token",
|
||||
authMode: "token",
|
||||
profileId: "profile-1",
|
||||
});
|
||||
const [streamModel, , streamOptions] = streamSimpleMock.mock.calls[0] ?? [];
|
||||
expectRecordFields(streamModel, {
|
||||
provider: "github-copilot",
|
||||
id: "gpt-5.4",
|
||||
baseUrl: "https://api.enterprise.githubcopilot.com",
|
||||
});
|
||||
expectRecordFields(streamOptions, { apiKey: "copilot-runtime-token" });
|
||||
});
|
||||
|
||||
it("uses the provider's stream fn when registered so provider URL construction runs (#68336)", async () => {
|
||||
@@ -540,17 +504,16 @@ describe("runBtwSideQuestion", () => {
|
||||
const result = await runSideQuestion({ provider: "ollama", model: "glm-5.1" });
|
||||
|
||||
expect(result).toEqual({ text: "Ollama Cloud answer." });
|
||||
const registerParams = expectRecordFields(
|
||||
registerProviderStreamForModelMock.mock.calls[0]?.[0],
|
||||
{
|
||||
expect(registerProviderStreamForModelMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
model: expect.objectContaining({
|
||||
provider: "ollama",
|
||||
api: "openai-completions",
|
||||
baseUrl: "https://ollama.com/",
|
||||
}),
|
||||
workspaceDir: "/tmp/workspace",
|
||||
},
|
||||
}),
|
||||
);
|
||||
expectRecordFields(registerParams.model, {
|
||||
provider: "ollama",
|
||||
api: "openai-completions",
|
||||
baseUrl: "https://ollama.com/",
|
||||
});
|
||||
expect(providerStreamFn).toHaveBeenCalledTimes(1);
|
||||
expect(streamSimpleMock).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -621,8 +584,16 @@ describe("runBtwSideQuestion", () => {
|
||||
const result = await runSideQuestion({ resolvedThinkLevel: "adaptive" });
|
||||
|
||||
expect(result).toEqual({ text: "Final answer." });
|
||||
const [, , options] = streamSimpleMock.mock.calls[0] ?? [];
|
||||
expect((options as { reasoning?: unknown } | undefined)?.reasoning).toBeUndefined();
|
||||
expect(streamSimpleMock).toHaveBeenCalledWith(
|
||||
expect.anything(),
|
||||
expect.anything(),
|
||||
expect.objectContaining({ reasoning: undefined }),
|
||||
);
|
||||
expect(streamSimpleMock).toHaveBeenCalledWith(
|
||||
expect.anything(),
|
||||
expect.anything(),
|
||||
expect.not.objectContaining({ reasoning: expect.anything() }),
|
||||
);
|
||||
});
|
||||
|
||||
it("fails when the current branch has no messages", async () => {
|
||||
@@ -651,19 +622,27 @@ describe("runBtwSideQuestion", () => {
|
||||
const result = await runMathSideQuestion();
|
||||
|
||||
expect(result).toEqual({ text: MATH_ANSWER });
|
||||
const context = streamContext();
|
||||
expect(String(context.systemPrompt)).toContain("ephemeral /btw side question");
|
||||
const messages = contextMessages(context);
|
||||
expect(messages.some((message) => message.role === "user")).toBe(true);
|
||||
const sideQuestionMessage = messages.find(
|
||||
(message) =>
|
||||
message.role === "user" &&
|
||||
firstTextBlockIncludes(
|
||||
message,
|
||||
`<btw_side_question>\n${MATH_QUESTION}\n</btw_side_question>`,
|
||||
),
|
||||
expect(streamSimpleMock).toHaveBeenCalledWith(
|
||||
expect.anything(),
|
||||
expect.objectContaining({
|
||||
systemPrompt: expect.stringContaining("ephemeral /btw side question"),
|
||||
messages: expect.arrayContaining([
|
||||
expect.objectContaining({ role: "user" }),
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content: [
|
||||
{
|
||||
type: "text",
|
||||
text: expect.stringContaining(
|
||||
`<btw_side_question>\n${MATH_QUESTION}\n</btw_side_question>`,
|
||||
),
|
||||
},
|
||||
],
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
expect.anything(),
|
||||
);
|
||||
expect(sideQuestionMessage).toBeDefined();
|
||||
});
|
||||
|
||||
it("uses the in-flight prompt as background only when there is no prior transcript context", async () => {
|
||||
@@ -678,11 +657,24 @@ describe("runBtwSideQuestion", () => {
|
||||
const result = await runSideQuestion({ question: "what are we doing?" });
|
||||
|
||||
expect(result).toEqual({ text: "You're building a tic-tac-toe game in Brainfuck." });
|
||||
const [message] = contextMessages(streamContext());
|
||||
expectRecordFields(message, { role: "user" });
|
||||
expectTextBlockContains(
|
||||
(message.content as Array<unknown>)[0],
|
||||
"<in_flight_main_task>\nbuild me a tic-tac-toe game in brainfuck\n</in_flight_main_task>",
|
||||
expect(streamSimpleMock).toHaveBeenCalledWith(
|
||||
expect.anything(),
|
||||
expect.objectContaining({
|
||||
messages: [
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content: [
|
||||
{
|
||||
type: "text",
|
||||
text: expect.stringContaining(
|
||||
"<in_flight_main_task>\nbuild me a tic-tac-toe game in brainfuck\n</in_flight_main_task>",
|
||||
),
|
||||
},
|
||||
],
|
||||
}),
|
||||
],
|
||||
}),
|
||||
expect.anything(),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -691,19 +683,27 @@ describe("runBtwSideQuestion", () => {
|
||||
|
||||
await runSideQuestion({ question: "what is the distance to the sun?" });
|
||||
|
||||
const context = streamContext();
|
||||
expect(String(context.systemPrompt)).toContain(
|
||||
"Do not continue, resume, or complete any unfinished task",
|
||||
);
|
||||
const sideQuestionMessage = contextMessages(context).find(
|
||||
(message) =>
|
||||
message.role === "user" &&
|
||||
firstTextBlockIncludes(
|
||||
message,
|
||||
"Ignore any unfinished task in the conversation while answering it.",
|
||||
),
|
||||
);
|
||||
expect(sideQuestionMessage).toBeDefined();
|
||||
const [, context] = streamSimpleMock.mock.calls[0] ?? [];
|
||||
expect(context).toMatchObject({
|
||||
systemPrompt: expect.stringContaining(
|
||||
"Do not continue, resume, or complete any unfinished task",
|
||||
),
|
||||
});
|
||||
expect(context).toMatchObject({
|
||||
messages: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content: [
|
||||
{
|
||||
type: "text",
|
||||
text: expect.stringContaining(
|
||||
"Ignore any unfinished task in the conversation while answering it.",
|
||||
),
|
||||
},
|
||||
],
|
||||
}),
|
||||
]),
|
||||
});
|
||||
});
|
||||
|
||||
it("branches away from an unresolved trailing user turn before building BTW context", async () => {
|
||||
@@ -816,12 +816,17 @@ describe("runBtwSideQuestion", () => {
|
||||
|
||||
await runMathSideQuestion();
|
||||
|
||||
const messages = contextMessages(streamContext());
|
||||
expect(messages).toHaveLength(3);
|
||||
expectRecordFields(messages[0], { role: "user" });
|
||||
expectRecordFields(messages[1], { role: "assistant" });
|
||||
expectRecordFields(messages[2], { role: "user" });
|
||||
expect(messages.some((message) => message.role === "toolResult")).toBe(false);
|
||||
const [, context] = streamSimpleMock.mock.calls[0] ?? [];
|
||||
expect(context).toMatchObject({
|
||||
messages: [
|
||||
expect.objectContaining({ role: "user" }),
|
||||
expect.objectContaining({ role: "assistant" }),
|
||||
expect.objectContaining({ role: "user" }),
|
||||
],
|
||||
});
|
||||
expect((context as { messages?: Array<{ role?: string }> }).messages).not.toEqual(
|
||||
expect.arrayContaining([expect.objectContaining({ role: "toolResult" })]),
|
||||
);
|
||||
});
|
||||
|
||||
it("strips assistant tool calls from BTW context so no-tool side questions stay tool-free", async () => {
|
||||
@@ -841,19 +846,23 @@ describe("runBtwSideQuestion", () => {
|
||||
|
||||
await runMathSideQuestion();
|
||||
|
||||
const context = streamContext();
|
||||
const [, context] = streamSimpleMock.mock.calls[0] ?? [];
|
||||
expectSanitizedAssistantContext(context, "Let me check.");
|
||||
const assistantMessages = contextMessages(context).filter(
|
||||
(message) => message.role === "assistant",
|
||||
expect(
|
||||
(context as { messages?: Array<{ role?: string; content?: Array<{ type?: string }> }> })
|
||||
.messages,
|
||||
).not.toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "assistant",
|
||||
content: expect.arrayContaining([
|
||||
expect.objectContaining({ type: "toolCall" }),
|
||||
expect.objectContaining({ type: "toolUse" }),
|
||||
expect.objectContaining({ type: "tool_call" }),
|
||||
]),
|
||||
}),
|
||||
]),
|
||||
);
|
||||
const assistantContentTypes = assistantMessages.flatMap((message) =>
|
||||
Array.isArray(message.content)
|
||||
? message.content.map((block) => (block as { type?: unknown }).type)
|
||||
: [],
|
||||
);
|
||||
expect(assistantContentTypes).not.toContain("toolCall");
|
||||
expect(assistantContentTypes).not.toContain("toolUse");
|
||||
expect(assistantContentTypes).not.toContain("tool_call");
|
||||
});
|
||||
|
||||
it("drops assistant messages that contain only tool calls", async () => {
|
||||
@@ -906,14 +915,17 @@ describe("runBtwSideQuestion", () => {
|
||||
const context = await runMathSideQuestionAndCaptureContext();
|
||||
|
||||
expectSanitizedAssistantContext(context, "Visible answer");
|
||||
const assistantContentTypes = contextMessages(context)
|
||||
.filter((message) => message.role === "assistant")
|
||||
.flatMap((message) =>
|
||||
Array.isArray(message.content)
|
||||
? message.content.map((block) => (block as { type?: unknown }).type)
|
||||
: [],
|
||||
);
|
||||
expect(assistantContentTypes).not.toContain("thinking");
|
||||
expect(
|
||||
(context as { messages?: Array<{ role?: string; content?: Array<{ type?: string }> }> })
|
||||
.messages,
|
||||
).not.toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "assistant",
|
||||
content: expect.arrayContaining([expect.objectContaining({ type: "thinking" })]),
|
||||
}),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
it("drops thinking-only assistant messages from BTW context", async () => {
|
||||
|
||||
@@ -157,50 +157,6 @@ function buildPreparedContext(params?: {
|
||||
};
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
expect(value, label).toBeTypeOf("object");
|
||||
expect(value, label).not.toBeNull();
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): Array<unknown> {
|
||||
expect(Array.isArray(value), label).toBe(true);
|
||||
return value as Array<unknown>;
|
||||
}
|
||||
|
||||
function callArg(
|
||||
mock: { mock: { calls: Array<Array<unknown>> } },
|
||||
callIndex: number,
|
||||
argIndex: number,
|
||||
label: string,
|
||||
) {
|
||||
const call = mock.mock.calls.at(callIndex);
|
||||
expect(call, label).toBeDefined();
|
||||
return call?.[argIndex];
|
||||
}
|
||||
|
||||
async function expectFailoverAttribution(
|
||||
run: Promise<unknown>,
|
||||
expected: { sessionId: string; lane: string },
|
||||
) {
|
||||
try {
|
||||
await run;
|
||||
throw new Error("expected run to fail");
|
||||
} catch (error) {
|
||||
const failure = requireRecord(error, "failover error");
|
||||
expect(failure.name).toBe("FailoverError");
|
||||
expect(failure.sessionId).toBe(expected.sessionId);
|
||||
expect(failure.lane).toBe(expected.lane);
|
||||
}
|
||||
}
|
||||
|
||||
function expectTextMessage(value: unknown, fields: { role: string; content: string }) {
|
||||
const message = requireRecord(value, "message");
|
||||
expect(message.role).toBe(fields.role);
|
||||
expect(message.content).toBe(fields.content);
|
||||
expect(message.timestamp).toBeTypeOf("number");
|
||||
}
|
||||
|
||||
describe("runCliAgent reliability", () => {
|
||||
afterEach(() => {
|
||||
replyRunTesting.resetReplyRunRegistry();
|
||||
@@ -245,7 +201,7 @@ describe("runCliAgent reliability", () => {
|
||||
}),
|
||||
);
|
||||
|
||||
await expectFailoverAttribution(
|
||||
await expect(
|
||||
executePreparedCliRun(
|
||||
buildPreparedContext({
|
||||
cliSessionId: "thread-123",
|
||||
@@ -254,8 +210,11 @@ describe("runCliAgent reliability", () => {
|
||||
}),
|
||||
"thread-123",
|
||||
),
|
||||
{ sessionId: "s1", lane: "custom-lane" },
|
||||
);
|
||||
).rejects.toMatchObject({
|
||||
name: "FailoverError",
|
||||
sessionId: "s1",
|
||||
lane: "custom-lane",
|
||||
});
|
||||
});
|
||||
|
||||
it("enqueues a system event and heartbeat wake on no-output watchdog timeout for session runs", async () => {
|
||||
@@ -287,7 +246,7 @@ describe("runCliAgent reliability", () => {
|
||||
const [notice, opts] = enqueueSystemEventMock.mock.calls[0] ?? [];
|
||||
expect(String(notice)).toContain("produced no output");
|
||||
expect(String(notice)).toContain("interactive input or an approval prompt");
|
||||
expect(requireRecord(opts, "system event options").sessionKey).toBe("agent:main:main");
|
||||
expect(opts).toMatchObject({ sessionKey: "agent:main:main" });
|
||||
expect(requestHeartbeatMock).toHaveBeenCalledWith({
|
||||
source: "cli-watchdog",
|
||||
intent: "event",
|
||||
@@ -381,17 +340,17 @@ describe("runCliAgent reliability", () => {
|
||||
expect(hookRunner.runLlmInput).toHaveBeenCalledTimes(1);
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
const agentEndEvent = requireRecord(
|
||||
callArg(hookRunner.runAgentEnd, 0, 0, "agent_end event"),
|
||||
"agent_end event",
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
success: false,
|
||||
error: "rate limit exceeded",
|
||||
messages: [
|
||||
{ role: "user", content: "earlier context", timestamp: expect.any(Number) },
|
||||
{ role: "user", content: "hi", timestamp: expect.any(Number) },
|
||||
],
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(agentEndEvent.success).toBe(false);
|
||||
expect(agentEndEvent.error).toBe("rate limit exceeded");
|
||||
const messages = requireArray(agentEndEvent.messages, "agent_end messages");
|
||||
expect(messages).toHaveLength(2);
|
||||
expectTextMessage(messages[0], { role: "user", content: "earlier context" });
|
||||
expectTextMessage(messages[1], { role: "user", content: "hi" });
|
||||
expect(callArg(hookRunner.runAgentEnd, 0, 1, "agent_end context")).toBeTypeOf("object");
|
||||
} finally {
|
||||
fs.rmSync(dir, { recursive: true, force: true });
|
||||
}
|
||||
@@ -419,20 +378,21 @@ describe("runCliAgent reliability", () => {
|
||||
expect(result.meta.finalPromptText).toContain("Warning: prompt budget low.");
|
||||
expect(result.meta.finalPromptText).toContain("hi");
|
||||
expect(result.meta.finalAssistantRawText).toBe("hello from cli");
|
||||
const executionTrace = requireRecord(result.meta.executionTrace, "execution trace");
|
||||
expect(executionTrace.winnerProvider).toBe("codex-cli");
|
||||
expect(executionTrace.winnerModel).toBe("gpt-5.4");
|
||||
expect(executionTrace.fallbackUsed).toBe(false);
|
||||
expect(executionTrace.runner).toBe("cli");
|
||||
expect(executionTrace.attempts).toEqual([
|
||||
{ provider: "codex-cli", model: "gpt-5.4", result: "success" },
|
||||
]);
|
||||
const requestShaping = requireRecord(result.meta.requestShaping, "request shaping");
|
||||
expect(requestShaping.thinking).toBe("low");
|
||||
const completion = requireRecord(result.meta.completion, "completion");
|
||||
expect(completion.finishReason).toBe("stop");
|
||||
expect(completion.stopReason).toBe("completed");
|
||||
expect(completion.refusal).toBe(false);
|
||||
expect(result.meta.executionTrace).toMatchObject({
|
||||
winnerProvider: "codex-cli",
|
||||
winnerModel: "gpt-5.4",
|
||||
fallbackUsed: false,
|
||||
runner: "cli",
|
||||
attempts: [{ provider: "codex-cli", model: "gpt-5.4", result: "success" }],
|
||||
});
|
||||
expect(result.meta.requestShaping).toMatchObject({
|
||||
thinking: "low",
|
||||
});
|
||||
expect(result.meta.completion).toMatchObject({
|
||||
finishReason: "stop",
|
||||
stopReason: "completed",
|
||||
refusal: false,
|
||||
});
|
||||
});
|
||||
|
||||
it("seeds fresh CLI sessions from the OpenClaw transcript", async () => {
|
||||
@@ -536,8 +496,7 @@ describe("runCliAgent reliability", () => {
|
||||
});
|
||||
|
||||
finishRun?.();
|
||||
const result = await run;
|
||||
expect(result.text).toBe("hello from cli");
|
||||
await expect(run).resolves.toMatchObject({ text: "hello from cli" });
|
||||
expect(replyRunRegistry.isStreaming("agent:main:main")).toBe(false);
|
||||
operation.complete();
|
||||
});
|
||||
@@ -617,60 +576,57 @@ describe("runCliAgent reliability", () => {
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
const llmInputEvent = requireRecord(
|
||||
callArg(hookRunner.runLlmInput, 0, 0, "llm_input event"),
|
||||
"llm_input event",
|
||||
expect(hookRunner.runLlmInput).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
runId: "run-2",
|
||||
sessionId: "s1",
|
||||
provider: "codex-cli",
|
||||
model: "gpt-5.4",
|
||||
prompt: "hi",
|
||||
systemPrompt: "You are a helpful assistant.",
|
||||
historyMessages: expect.any(Array),
|
||||
imagesCount: 0,
|
||||
}),
|
||||
expect.objectContaining({
|
||||
runId: "run-2",
|
||||
agentId: "main",
|
||||
sessionKey: "agent:main:main",
|
||||
sessionId: "s1",
|
||||
workspaceDir: dir,
|
||||
messageProvider: "acp",
|
||||
trigger: "user",
|
||||
channelId: "telegram",
|
||||
}),
|
||||
);
|
||||
expect(llmInputEvent.runId).toBe("run-2");
|
||||
expect(llmInputEvent.sessionId).toBe("s1");
|
||||
expect(llmInputEvent.provider).toBe("codex-cli");
|
||||
expect(llmInputEvent.model).toBe("gpt-5.4");
|
||||
expect(llmInputEvent.prompt).toBe("hi");
|
||||
expect(llmInputEvent.systemPrompt).toBe("You are a helpful assistant.");
|
||||
expect(Array.isArray(llmInputEvent.historyMessages)).toBe(true);
|
||||
expect(llmInputEvent.imagesCount).toBe(0);
|
||||
|
||||
const llmInputContext = requireRecord(
|
||||
callArg(hookRunner.runLlmInput, 0, 1, "llm_input context"),
|
||||
"llm_input context",
|
||||
expect(hookRunner.runLlmOutput).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
runId: "run-2",
|
||||
sessionId: "s1",
|
||||
provider: "codex-cli",
|
||||
model: "gpt-5.4",
|
||||
assistantTexts: ["hello from cli"],
|
||||
lastAssistant: expect.objectContaining({
|
||||
role: "assistant",
|
||||
content: [{ type: "text", text: "hello from cli" }],
|
||||
provider: "codex-cli",
|
||||
model: "gpt-5.4",
|
||||
}),
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(llmInputContext.runId).toBe("run-2");
|
||||
expect(llmInputContext.agentId).toBe("main");
|
||||
expect(llmInputContext.sessionKey).toBe("agent:main:main");
|
||||
expect(llmInputContext.sessionId).toBe("s1");
|
||||
expect(llmInputContext.workspaceDir).toBe(dir);
|
||||
expect(llmInputContext.messageProvider).toBe("acp");
|
||||
expect(llmInputContext.trigger).toBe("user");
|
||||
expect(llmInputContext.channelId).toBe("telegram");
|
||||
|
||||
const llmOutputEvent = requireRecord(
|
||||
callArg(hookRunner.runLlmOutput, 0, 0, "llm_output event"),
|
||||
"llm_output event",
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
success: true,
|
||||
messages: [
|
||||
{ role: "user", content: "hi", timestamp: expect.any(Number) },
|
||||
expect.objectContaining({
|
||||
role: "assistant",
|
||||
content: [{ type: "text", text: "hello from cli" }],
|
||||
}),
|
||||
],
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(llmOutputEvent.runId).toBe("run-2");
|
||||
expect(llmOutputEvent.sessionId).toBe("s1");
|
||||
expect(llmOutputEvent.provider).toBe("codex-cli");
|
||||
expect(llmOutputEvent.model).toBe("gpt-5.4");
|
||||
expect(llmOutputEvent.assistantTexts).toEqual(["hello from cli"]);
|
||||
const lastAssistant = requireRecord(llmOutputEvent.lastAssistant, "last assistant");
|
||||
expect(lastAssistant.role).toBe("assistant");
|
||||
expect(lastAssistant.content).toEqual([{ type: "text", text: "hello from cli" }]);
|
||||
expect(lastAssistant.provider).toBe("codex-cli");
|
||||
expect(lastAssistant.model).toBe("gpt-5.4");
|
||||
expect(callArg(hookRunner.runLlmOutput, 0, 1, "llm_output context")).toBeTypeOf("object");
|
||||
|
||||
const agentEndEvent = requireRecord(
|
||||
callArg(hookRunner.runAgentEnd, 0, 0, "agent_end event"),
|
||||
"agent_end event",
|
||||
);
|
||||
expect(agentEndEvent.success).toBe(true);
|
||||
const messages = requireArray(agentEndEvent.messages, "agent_end messages");
|
||||
expect(messages).toHaveLength(2);
|
||||
expectTextMessage(messages[0], { role: "user", content: "hi" });
|
||||
const assistantMessage = requireRecord(messages[1], "assistant message");
|
||||
expect(assistantMessage.role).toBe("assistant");
|
||||
expect(assistantMessage.content).toEqual([{ type: "text", text: "hello from cli" }]);
|
||||
expect(callArg(hookRunner.runAgentEnd, 0, 1, "agent_end context")).toBeTypeOf("object");
|
||||
} finally {
|
||||
fs.rmSync(dir, { recursive: true, force: true });
|
||||
}
|
||||
@@ -720,48 +676,37 @@ describe("runCliAgent reliability", () => {
|
||||
expect(result.meta.livenessState).toBe("blocked");
|
||||
expect(supervisorSpawnMock).not.toHaveBeenCalled();
|
||||
expect(hookRunner.runLlmInput).not.toHaveBeenCalled();
|
||||
const beforeRunEvent = requireRecord(
|
||||
callArg(hookRunner.runBeforeAgentRun, 0, 0, "before_agent_run event"),
|
||||
"before_agent_run event",
|
||||
);
|
||||
expect(beforeRunEvent.prompt).toBe("secret prompt");
|
||||
const beforeRunMessages = requireArray(beforeRunEvent.messages, "before_agent_run messages");
|
||||
expect(
|
||||
beforeRunMessages.some((message) => {
|
||||
const record = requireRecord(message, "before_agent_run message");
|
||||
return record.role === "user" && record.content === "earlier context";
|
||||
expect(hookRunner.runBeforeAgentRun).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
prompt: "secret prompt",
|
||||
messages: expect.arrayContaining([
|
||||
expect.objectContaining({ role: "user", content: "earlier context" }),
|
||||
]),
|
||||
}),
|
||||
expect.objectContaining({
|
||||
runId: "run-blocked-cli",
|
||||
agentId: "main",
|
||||
sessionKey: "agent:main:main",
|
||||
}),
|
||||
).toBe(true);
|
||||
const beforeRunContext = requireRecord(
|
||||
callArg(hookRunner.runBeforeAgentRun, 0, 1, "before_agent_run context"),
|
||||
"before_agent_run context",
|
||||
);
|
||||
expect(beforeRunContext.runId).toBe("run-blocked-cli");
|
||||
expect(beforeRunContext.agentId).toBe("main");
|
||||
expect(beforeRunContext.sessionKey).toBe("agent:main:main");
|
||||
await vi.waitFor(() => {
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
const agentEndEvent = requireRecord(
|
||||
callArg(hookRunner.runAgentEnd, 0, 0, "agent_end event"),
|
||||
"agent_end event",
|
||||
);
|
||||
expect(agentEndEvent.success).toBe(false);
|
||||
expect(agentEndEvent.error).toBe(
|
||||
"Your message could not be sent: The agent cannot read this message. (blocked by policy-plugin)",
|
||||
);
|
||||
const agentEndMessages = requireArray(agentEndEvent.messages, "agent_end messages");
|
||||
expect(
|
||||
agentEndMessages.some((message) => {
|
||||
const record = requireRecord(message, "agent_end message");
|
||||
return (
|
||||
record.role === "user" &&
|
||||
record.content ===
|
||||
"Your message could not be sent: The agent cannot read this message. (blocked by policy-plugin)"
|
||||
);
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
success: false,
|
||||
error:
|
||||
"Your message could not be sent: The agent cannot read this message. (blocked by policy-plugin)",
|
||||
messages: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
role: "user",
|
||||
content:
|
||||
"Your message could not be sent: The agent cannot read this message. (blocked by policy-plugin)",
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
).toBe(true);
|
||||
expect(callArg(hookRunner.runAgentEnd, 0, 1, "agent_end context")).toBeTypeOf("object");
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(JSON.stringify(hookRunner.runAgentEnd.mock.calls)).not.toContain("secret prompt");
|
||||
|
||||
const lines = fs.readFileSync(sessionFile, "utf-8").trim().split("\n");
|
||||
@@ -771,7 +716,9 @@ describe("runCliAgent reliability", () => {
|
||||
);
|
||||
expect(JSON.stringify(blockedLine)).not.toContain("secret prompt");
|
||||
expect(JSON.stringify(blockedLine)).not.toContain("matched secret prompt");
|
||||
expect(blockedLine.message.__openclaw.beforeAgentRunBlocked.blockedBy).toBe("policy-plugin");
|
||||
expect(blockedLine.message.__openclaw.beforeAgentRunBlocked).toMatchObject({
|
||||
blockedBy: "policy-plugin",
|
||||
});
|
||||
expect(blockedLine.message.__openclaw.beforeAgentRunBlocked).not.toHaveProperty("reason");
|
||||
expect(Object.hasOwn(blockedLine.message.__openclaw, "beforeAgentRunBlocked")).toBe(true);
|
||||
} finally {
|
||||
@@ -839,16 +786,14 @@ describe("runCliAgent reliability", () => {
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
const agentEndEvent = requireRecord(
|
||||
callArg(hookRunner.runAgentEnd, 0, 0, "agent_end event"),
|
||||
"agent_end event",
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
success: false,
|
||||
error: "rate limit exceeded",
|
||||
messages: [{ role: "user", content: "hi", timestamp: expect.any(Number) }],
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(agentEndEvent.success).toBe(false);
|
||||
expect(agentEndEvent.error).toBe("rate limit exceeded");
|
||||
const messages = requireArray(agentEndEvent.messages, "agent_end messages");
|
||||
expect(messages).toHaveLength(1);
|
||||
expectTextMessage(messages[0], { role: "user", content: "hi" });
|
||||
expect(callArg(hookRunner.runAgentEnd, 0, 1, "agent_end context")).toBeTypeOf("object");
|
||||
});
|
||||
|
||||
it("does not emit duplicate llm_input when session-expired recovery succeeds", async () => {
|
||||
@@ -916,7 +861,9 @@ describe("runCliAgent reliability", () => {
|
||||
},
|
||||
});
|
||||
|
||||
expect(result.payloads).toEqual([{ text: "recovered output" }]);
|
||||
expect(result).toMatchObject({
|
||||
payloads: [{ text: "recovered output" }],
|
||||
});
|
||||
expect(result.meta.finalPromptText).toContain("User: recovered history");
|
||||
|
||||
await vi.waitFor(() => {
|
||||
@@ -924,15 +871,14 @@ describe("runCliAgent reliability", () => {
|
||||
expect(hookRunner.runLlmOutput).toHaveBeenCalledTimes(1);
|
||||
expect(hookRunner.runAgentEnd).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
const llmInputEvent = requireRecord(
|
||||
callArg(hookRunner.runLlmInput, 0, 0, "llm_input event"),
|
||||
"llm_input event",
|
||||
);
|
||||
const historyMessages = requireArray(llmInputEvent.historyMessages, "history messages");
|
||||
expect(historyMessages).toHaveLength(MAX_CLI_SESSION_HISTORY_MESSAGES);
|
||||
const firstHistoryMessage = requireRecord(historyMessages[0], "first history message");
|
||||
expect(firstHistoryMessage.role).toBe("user");
|
||||
expect(firstHistoryMessage.content).toBe(`history-5`);
|
||||
const llmInputCalls = hookRunner.runLlmInput.mock.calls as unknown as Array<Array<unknown>>;
|
||||
const llmInputEvent = llmInputCalls[0]?.[0] as { historyMessages: unknown[] } | undefined;
|
||||
expect(llmInputEvent).toMatchObject({ historyMessages: expect.any(Array) });
|
||||
expect(llmInputEvent?.historyMessages).toHaveLength(MAX_CLI_SESSION_HISTORY_MESSAGES);
|
||||
expect(llmInputEvent?.historyMessages[0]).toMatchObject({
|
||||
role: "user",
|
||||
content: `history-5`,
|
||||
});
|
||||
} finally {
|
||||
fs.rmSync(dir, { recursive: true, force: true });
|
||||
}
|
||||
|
||||
@@ -188,64 +188,6 @@ describe("models-config", () => {
|
||||
expect(observedSnapshot).toBe(pluginMetadataSnapshot);
|
||||
});
|
||||
|
||||
it("normalizes retired Gemini ids preserved from existing models.json rows", async () => {
|
||||
const plan = await planOpenClawModelsJsonWithDeps(
|
||||
{
|
||||
cfg: { models: { mode: "merge", providers: {} } },
|
||||
agentDir: "/tmp/openclaw-models-config-env-vars-test",
|
||||
env: {},
|
||||
existingRaw: "",
|
||||
existingParsed: {
|
||||
providers: {
|
||||
google: {
|
||||
baseUrl: "https://generativelanguage.googleapis.com/v1beta",
|
||||
api: "google-generative-ai",
|
||||
apiKey: "GOOGLE_API_KEY", // pragma: allowlist secret
|
||||
models: [
|
||||
{
|
||||
id: "gemini-3-pro-preview",
|
||||
name: "Gemini 3 Pro",
|
||||
input: ["text"],
|
||||
},
|
||||
],
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
resolveImplicitProviders: async () => ({
|
||||
openai: {
|
||||
baseUrl: "https://api.openai.com/v1",
|
||||
api: "openai-responses",
|
||||
apiKey: "OPENAI_API_KEY", // pragma: allowlist secret
|
||||
models: [
|
||||
{
|
||||
id: "gpt-5.5",
|
||||
name: "GPT-5.5",
|
||||
input: ["text"],
|
||||
reasoning: true,
|
||||
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
|
||||
contextWindow: 400000,
|
||||
maxTokens: 128000,
|
||||
},
|
||||
],
|
||||
},
|
||||
}),
|
||||
},
|
||||
);
|
||||
|
||||
expect(plan.action).toBe("write");
|
||||
if (plan.action !== "write") {
|
||||
throw new Error("Expected models.json write plan");
|
||||
}
|
||||
const parsed = JSON.parse(plan.contents) as {
|
||||
providers?: Record<string, { models?: Array<{ id?: string }> }>;
|
||||
};
|
||||
expect(parsed.providers?.google?.models?.map((model) => model.id)).toEqual([
|
||||
"gemini-3.1-pro-preview",
|
||||
]);
|
||||
});
|
||||
|
||||
it("uses config env.vars entries for implicit provider discovery without mutating process.env", async () => {
|
||||
await withTempEnv(["OPENROUTER_API_KEY", TEST_ENV_VAR], async () => {
|
||||
unsetEnv(["OPENROUTER_API_KEY", TEST_ENV_VAR]);
|
||||
|
||||
@@ -9,7 +9,6 @@ import {
|
||||
import {
|
||||
applyNativeStreamingUsageCompat,
|
||||
enforceSourceManagedProviderSecrets,
|
||||
normalizeProviderCatalogModelsForConfig,
|
||||
normalizeProviders,
|
||||
resolveImplicitProviders,
|
||||
type ProviderConfig,
|
||||
@@ -168,15 +167,13 @@ export async function planOpenClawModelsJsonWithDeps(
|
||||
providers: normalizedProviders,
|
||||
secretRefManagedProviders,
|
||||
});
|
||||
const normalizedMergedProviders =
|
||||
normalizeProviderCatalogModelsForConfig(mergedProviders) ?? mergedProviders;
|
||||
const secretEnforcedProviders =
|
||||
enforceSourceManagedProviderSecrets({
|
||||
providers: normalizedMergedProviders,
|
||||
providers: mergedProviders,
|
||||
sourceProviders: params.sourceConfigForSecrets?.models?.providers,
|
||||
sourceSecretDefaults: params.sourceConfigForSecrets?.secrets?.defaults,
|
||||
secretRefManagedProviders,
|
||||
}) ?? normalizedMergedProviders;
|
||||
}) ?? mergedProviders;
|
||||
const finalProviders = applyNativeStreamingUsageCompat(secretEnforcedProviders);
|
||||
const nextContents = `${JSON.stringify({ providers: finalProviders }, null, 2)}\n`;
|
||||
|
||||
|
||||
@@ -84,26 +84,6 @@ function normalizeProviderModelsForConfig(
|
||||
: { provider, mutated };
|
||||
}
|
||||
|
||||
export function normalizeProviderCatalogModelsForConfig(
|
||||
providers: ModelsConfig["providers"],
|
||||
): ModelsConfig["providers"] {
|
||||
if (!providers) {
|
||||
return providers;
|
||||
}
|
||||
|
||||
let mutated = false;
|
||||
const next: Record<string, ProviderConfig> = {};
|
||||
for (const [providerKey, provider] of Object.entries(providers)) {
|
||||
const normalized = normalizeProviderModelsForConfig(providerKey, provider);
|
||||
if (normalized.mutated) {
|
||||
mutated = true;
|
||||
}
|
||||
next[providerKey] = normalized.provider;
|
||||
}
|
||||
|
||||
return mutated ? next : providers;
|
||||
}
|
||||
|
||||
export function normalizeProviders(params: {
|
||||
providers: ModelsConfig["providers"];
|
||||
agentDir: string;
|
||||
|
||||
@@ -1,8 +1,5 @@
|
||||
export { resolveImplicitProviders } from "./models-config.providers.implicit.js";
|
||||
export {
|
||||
normalizeProviderCatalogModelsForConfig,
|
||||
normalizeProviders,
|
||||
} from "./models-config.providers.normalize.js";
|
||||
export { normalizeProviders } from "./models-config.providers.normalize.js";
|
||||
export type { ProviderConfig } from "./models-config.providers.secrets.js";
|
||||
export { applyNativeStreamingUsageCompat } from "./models-config.providers.policy.js";
|
||||
export { enforceSourceManagedProviderSecrets } from "./models-config.providers.source-managed.js";
|
||||
|
||||
@@ -9,7 +9,6 @@ import type {
|
||||
import { sleepWithAbort } from "../../infra/backoff.js";
|
||||
import { formatErrorMessage } from "../../infra/errors.js";
|
||||
import { enqueueCommandInLane, getQueueSize } from "../../process/command-queue.js";
|
||||
import { CommandPriority } from "../../process/command-queue.types.js";
|
||||
import { normalizeOptionalString } from "../../shared/string-coerce.js";
|
||||
import {
|
||||
completeTaskRunByRunId,
|
||||
@@ -316,7 +315,6 @@ export function buildContextEngineMaintenanceRuntimeContext(params: {
|
||||
return await enqueueCommandInLane(
|
||||
resolveSessionLane(rewriteSessionKey),
|
||||
async () => await rewriteTranscriptEntriesInFile(),
|
||||
{ priority: CommandPriority.Low },
|
||||
);
|
||||
}
|
||||
return await rewriteTranscriptEntriesInFile();
|
||||
@@ -571,21 +569,18 @@ function scheduleDeferredTurnMaintenance(params: DeferredTurnMaintenanceSchedule
|
||||
const schedulerAbort = createDeferredTurnMaintenanceAbortSignal();
|
||||
let runPromise: Promise<void>;
|
||||
try {
|
||||
runPromise = enqueueCommandInLane(
|
||||
resolveDeferredTurnMaintenanceLane(sessionKey),
|
||||
async () =>
|
||||
runDeferredTurnMaintenanceWorker({
|
||||
contextEngine: params.contextEngine,
|
||||
sessionId: params.sessionId,
|
||||
sessionKey,
|
||||
sessionFile: params.sessionFile,
|
||||
sessionManager: params.sessionManager,
|
||||
runtimeContext: params.runtimeContext,
|
||||
agentId: params.agentId,
|
||||
config: params.config,
|
||||
runId: task.runId!,
|
||||
}),
|
||||
{ priority: CommandPriority.Low },
|
||||
runPromise = enqueueCommandInLane(resolveDeferredTurnMaintenanceLane(sessionKey), async () =>
|
||||
runDeferredTurnMaintenanceWorker({
|
||||
contextEngine: params.contextEngine,
|
||||
sessionId: params.sessionId,
|
||||
sessionKey,
|
||||
sessionFile: params.sessionFile,
|
||||
sessionManager: params.sessionManager,
|
||||
runtimeContext: params.runtimeContext,
|
||||
agentId: params.agentId,
|
||||
config: params.config,
|
||||
runId: task.runId!,
|
||||
}),
|
||||
);
|
||||
} catch (err) {
|
||||
schedulerAbort.dispose();
|
||||
|
||||
@@ -16,10 +16,7 @@ import { buildAgentHookContextChannelFields } from "../../plugins/hook-agent-con
|
||||
import { getGlobalHookRunner } from "../../plugins/hook-runner-global.js";
|
||||
import { resolveProviderAuthProfileId } from "../../plugins/provider-runtime.js";
|
||||
import { enqueueCommandInLane } from "../../process/command-queue.js";
|
||||
import {
|
||||
CommandPriority,
|
||||
type CommandQueueEnqueueOptions,
|
||||
} from "../../process/command-queue.types.js";
|
||||
import type { CommandQueueEnqueueOptions } from "../../process/command-queue.types.js";
|
||||
import { normalizeOptionalString } from "../../shared/string-coerce.js";
|
||||
import { sanitizeForLog } from "../../terminal/ansi.js";
|
||||
import { resolveUserPath } from "../../utils.js";
|
||||
@@ -219,23 +216,6 @@ function withEmbeddedRunLaneTimeout(
|
||||
return { ...opts, taskTimeoutMs: laneTaskTimeoutMs };
|
||||
}
|
||||
|
||||
function resolveEmbeddedRunQueuePriority(params: RunEmbeddedPiAgentParams): CommandPriority {
|
||||
if (params.queuePriority !== undefined) {
|
||||
return params.queuePriority;
|
||||
}
|
||||
switch (params.trigger) {
|
||||
case "user":
|
||||
case "manual":
|
||||
return CommandPriority.High;
|
||||
case "cron":
|
||||
case "heartbeat":
|
||||
case "memory":
|
||||
return CommandPriority.Low;
|
||||
default:
|
||||
return CommandPriority.Normal;
|
||||
}
|
||||
}
|
||||
|
||||
function normalizeEmbeddedRunAttemptResult(
|
||||
attempt: EmbeddedRunAttemptForRunner,
|
||||
): EmbeddedRunAttemptForRunner {
|
||||
@@ -395,21 +375,14 @@ export async function runEmbeddedPiAgent(
|
||||
const sessionLane = resolveSessionLane(params.sessionKey?.trim() || params.sessionId);
|
||||
const globalLane = resolveGlobalLane(params.lane);
|
||||
const laneTaskTimeoutMs = resolveEmbeddedRunLaneTimeoutMs(params.timeoutMs);
|
||||
const queuePriority = resolveEmbeddedRunQueuePriority(params);
|
||||
const withQueuePriority = (opts?: CommandQueueEnqueueOptions): CommandQueueEnqueueOptions => ({
|
||||
...opts,
|
||||
priority: opts?.priority ?? queuePriority,
|
||||
});
|
||||
const withLaneTimeout = (opts?: CommandQueueEnqueueOptions) =>
|
||||
withEmbeddedRunLaneTimeout(withQueuePriority(opts), laneTaskTimeoutMs);
|
||||
withEmbeddedRunLaneTimeout(opts, laneTaskTimeoutMs);
|
||||
const enqueueGlobal = <T>(task: () => Promise<T>, opts?: CommandQueueEnqueueOptions) =>
|
||||
params.enqueue
|
||||
? params.enqueue(task, withLaneTimeout(opts))
|
||||
: enqueueCommandInLane(globalLane, task, withLaneTimeout(opts));
|
||||
const enqueueSession = <T>(task: () => Promise<T>, opts?: CommandQueueEnqueueOptions) =>
|
||||
params.enqueue
|
||||
? params.enqueue(task, withQueuePriority(opts))
|
||||
: enqueueCommandInLane(sessionLane, task, withQueuePriority(opts));
|
||||
params.enqueue ? params.enqueue(task, opts) : enqueueCommandInLane(sessionLane, task, opts);
|
||||
const channelHint = params.messageChannel ?? params.messageProvider;
|
||||
const resolvedToolResultFormat =
|
||||
params.toolResultFormat ??
|
||||
|
||||
@@ -95,8 +95,6 @@ export function createSubscriptionMock(): SubscriptionMock {
|
||||
return {
|
||||
assistantTexts: [] as string[],
|
||||
toolMetas: [] as Array<{ toolName: string; meta?: string }>,
|
||||
runToolLifecycle: async <T>(toolParams: { execute: () => Promise<T> }) =>
|
||||
await toolParams.execute(),
|
||||
unsubscribe: () => {},
|
||||
setTerminalLifecycleMeta: () => {},
|
||||
waitForCompactionRetry: async () => {},
|
||||
|
||||
@@ -12,9 +12,6 @@ import { buildAgentSystemPrompt } from "../../system-prompt.js";
|
||||
import { resolveBootstrapContextTargets } from "./attempt-bootstrap-routing.js";
|
||||
import {
|
||||
buildContextEnginePromptCacheInfo,
|
||||
buildAutoAddedToolSearchControlNamesForAllowlistCheck,
|
||||
buildCallableToolNamesForEmptyAllowlistCheck,
|
||||
buildToolSearchRunPlan,
|
||||
buildAfterTurnRuntimeContext,
|
||||
buildAfterTurnRuntimeContextFromUsage,
|
||||
composeSystemPromptWithHookContext,
|
||||
@@ -91,137 +88,6 @@ describe("buildEmbeddedAttemptToolRunContext", () => {
|
||||
});
|
||||
});
|
||||
|
||||
describe("buildCallableToolNamesForEmptyAllowlistCheck", () => {
|
||||
it("ignores auto-added Tool Search controls so bad allowlists still fail", () => {
|
||||
expect(
|
||||
buildCallableToolNamesForEmptyAllowlistCheck({
|
||||
effectiveToolNames: ["tool_search_code"],
|
||||
autoAddedToolSearchControlNames: new Set(["tool_search_code"]),
|
||||
toolSearchCatalogToolCount: 0,
|
||||
}),
|
||||
).toEqual([]);
|
||||
});
|
||||
|
||||
it("counts cataloged tools hidden behind auto-added Tool Search controls", () => {
|
||||
expect(
|
||||
buildCallableToolNamesForEmptyAllowlistCheck({
|
||||
effectiveToolNames: ["tool_search_code"],
|
||||
autoAddedToolSearchControlNames: new Set(["tool_search_code"]),
|
||||
toolSearchCatalogToolCount: 1,
|
||||
}),
|
||||
).toEqual(["tool-search:0"]);
|
||||
});
|
||||
|
||||
it("keeps explicitly requested Tool Search controls callable", () => {
|
||||
expect(
|
||||
buildCallableToolNamesForEmptyAllowlistCheck({
|
||||
effectiveToolNames: ["tool_search_code"],
|
||||
autoAddedToolSearchControlNames: new Set(),
|
||||
toolSearchCatalogToolCount: 0,
|
||||
}),
|
||||
).toEqual(["tool_search_code"]);
|
||||
});
|
||||
});
|
||||
|
||||
describe("buildAutoAddedToolSearchControlNamesForAllowlistCheck", () => {
|
||||
it("treats controls as auto-added unless any explicit allowlist requested them", () => {
|
||||
expect(
|
||||
buildAutoAddedToolSearchControlNamesForAllowlistCheck({
|
||||
toolSearchControlsEnabled: true,
|
||||
explicitAllowlistSources: [{ entries: ["missing_tool"] }],
|
||||
controlNames: ["tool_search_code", "tool_search"],
|
||||
}),
|
||||
).toEqual(new Set(["tool_search_code", "tool_search"]));
|
||||
|
||||
expect(
|
||||
buildAutoAddedToolSearchControlNamesForAllowlistCheck({
|
||||
toolSearchControlsEnabled: true,
|
||||
explicitAllowlistSources: [{ entries: ["tool_search_code"] }],
|
||||
controlNames: ["tool_search_code", "tool_search"],
|
||||
}),
|
||||
).toEqual(new Set(["tool_search"]));
|
||||
});
|
||||
});
|
||||
|
||||
describe("buildToolSearchRunPlan", () => {
|
||||
it("keeps compact visible names separate from replay-safe names", () => {
|
||||
const plan = buildToolSearchRunPlan({
|
||||
visibleTools: [{ name: "tool_search_code" }] as never,
|
||||
uncompactedTools: [
|
||||
{ name: "tool_search_code" },
|
||||
{ name: "exec" },
|
||||
{ name: "fake_plugin_tool" },
|
||||
] as never,
|
||||
clientTools: [
|
||||
{
|
||||
type: "function",
|
||||
function: {
|
||||
name: "client_pick_file",
|
||||
parameters: { type: "object", properties: {} },
|
||||
},
|
||||
},
|
||||
],
|
||||
catalogRegistered: true,
|
||||
catalogToolCount: 2,
|
||||
controlsEnabled: true,
|
||||
explicitAllowlistSources: [{ entries: ["missing_tool"] }],
|
||||
});
|
||||
|
||||
expect([...plan.visibleAllowedToolNames]).toEqual(["tool_search_code"]);
|
||||
expect([...plan.replayAllowedToolNames]).toEqual([
|
||||
"tool_search_code",
|
||||
"exec",
|
||||
"fake_plugin_tool",
|
||||
"client_pick_file",
|
||||
]);
|
||||
expect(plan.emptyAllowlistCallableNames).toEqual(["tool-search:0", "tool-search:1"]);
|
||||
});
|
||||
|
||||
it("counts explicitly allowlisted client tools before they are cataloged later", () => {
|
||||
const plan = buildToolSearchRunPlan({
|
||||
visibleTools: [{ name: "tool_search_code" }] as never,
|
||||
uncompactedTools: [{ name: "tool_search_code" }] as never,
|
||||
clientTools: [
|
||||
{
|
||||
type: "function",
|
||||
function: {
|
||||
name: "client_pick_file",
|
||||
parameters: { type: "object", properties: {} },
|
||||
},
|
||||
},
|
||||
],
|
||||
catalogRegistered: true,
|
||||
catalogToolCount: 0,
|
||||
controlsEnabled: true,
|
||||
explicitAllowlistSources: [{ entries: ["client_pick_file"] }],
|
||||
});
|
||||
|
||||
expect(plan.emptyAllowlistCallableNames).toEqual(["tool-search-client:client_pick_file"]);
|
||||
});
|
||||
|
||||
it("does not let unrelated client tools mask a bad explicit allowlist", () => {
|
||||
const plan = buildToolSearchRunPlan({
|
||||
visibleTools: [{ name: "tool_search_code" }] as never,
|
||||
uncompactedTools: [{ name: "tool_search_code" }] as never,
|
||||
clientTools: [
|
||||
{
|
||||
type: "function",
|
||||
function: {
|
||||
name: "client_pick_file",
|
||||
parameters: { type: "object", properties: {} },
|
||||
},
|
||||
},
|
||||
],
|
||||
catalogRegistered: true,
|
||||
catalogToolCount: 0,
|
||||
controlsEnabled: true,
|
||||
explicitAllowlistSources: [{ entries: ["missing_tool"] }],
|
||||
});
|
||||
|
||||
expect(plan.emptyAllowlistCallableNames).toEqual([]);
|
||||
});
|
||||
});
|
||||
|
||||
describe("normalizeMessagesForLlmBoundary", () => {
|
||||
it("strips tool result details before provider conversion", () => {
|
||||
const input = [
|
||||
|
||||
@@ -174,22 +174,6 @@ import {
|
||||
collectExplicitToolAllowlistSources,
|
||||
} from "../../tool-allowlist-guard.js";
|
||||
import { UNKNOWN_TOOL_THRESHOLD } from "../../tool-loop-detection.js";
|
||||
import { normalizeToolName } from "../../tool-policy.js";
|
||||
import {
|
||||
addClientToolsToToolSearchCatalog,
|
||||
applyToolSearchCatalog,
|
||||
clearToolSearchCatalog,
|
||||
createToolSearchCatalogRef,
|
||||
projectToolSearchTargetTranscriptMessages,
|
||||
resolveToolSearchConfig,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
type ToolSearchCatalogRef,
|
||||
type ToolSearchCatalogToolExecutor,
|
||||
type ToolSearchTargetTranscriptProjection,
|
||||
} from "../../tool-search.js";
|
||||
import { shouldAllowProviderOwnedThinkingReplay } from "../../transcript-policy.js";
|
||||
import { normalizeUsage, type NormalizedUsage } from "../../usage.js";
|
||||
import { DEFAULT_BOOTSTRAP_FILENAME } from "../../workspace.js";
|
||||
@@ -244,7 +228,6 @@ import { applySystemPromptOverrideToSession } from "../system-prompt.js";
|
||||
import { dropReasoningFromHistory, dropThinkingBlocks } from "../thinking.js";
|
||||
import {
|
||||
collectAllowedToolNames,
|
||||
collectCoreBuiltinToolNames,
|
||||
collectRegisteredToolNames,
|
||||
PI_RESERVED_TOOL_NAMES,
|
||||
toSessionToolAllowlist,
|
||||
@@ -406,114 +389,6 @@ export {
|
||||
};
|
||||
|
||||
const MAX_BTW_SNAPSHOT_MESSAGES = 100;
|
||||
const TOOL_SEARCH_CONTROL_ALLOWLIST_NAMES = [
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
];
|
||||
|
||||
export function buildCallableToolNamesForEmptyAllowlistCheck(params: {
|
||||
effectiveToolNames: string[];
|
||||
autoAddedToolSearchControlNames?: Set<string>;
|
||||
toolSearchCatalogToolCount: number;
|
||||
}): string[] {
|
||||
return [
|
||||
...params.effectiveToolNames.filter(
|
||||
(toolName) => !params.autoAddedToolSearchControlNames?.has(toolName),
|
||||
),
|
||||
...Array.from(
|
||||
{ length: params.toolSearchCatalogToolCount },
|
||||
(_, index) => `tool-search:${index}`,
|
||||
),
|
||||
];
|
||||
}
|
||||
|
||||
export function buildAutoAddedToolSearchControlNamesForAllowlistCheck(params: {
|
||||
toolSearchControlsEnabled: boolean;
|
||||
explicitAllowlistSources: Array<{ entries: string[] }>;
|
||||
controlNames?: readonly string[];
|
||||
}): Set<string> | undefined {
|
||||
if (!params.toolSearchControlsEnabled) {
|
||||
return undefined;
|
||||
}
|
||||
const explicitlyAllowed = new Set(
|
||||
params.explicitAllowlistSources.flatMap((source) =>
|
||||
source.entries.map((entry) => normalizeToolName(entry)),
|
||||
),
|
||||
);
|
||||
return new Set(
|
||||
(params.controlNames ?? TOOL_SEARCH_CONTROL_ALLOWLIST_NAMES).filter(
|
||||
(controlName) => !explicitlyAllowed.has(normalizeToolName(controlName)),
|
||||
),
|
||||
);
|
||||
}
|
||||
|
||||
export type ToolSearchRunPlan = {
|
||||
visibleAllowedToolNames: Set<string>;
|
||||
replayAllowedToolNames: Set<string>;
|
||||
autoAddedControlNames?: Set<string>;
|
||||
emptyAllowlistCallableNames: string[];
|
||||
};
|
||||
|
||||
type CollectAllowedToolNamesParams = Parameters<typeof collectAllowedToolNames>[0];
|
||||
|
||||
function collectExplicitlyAllowedClientToolNames(params: {
|
||||
clientTools?: CollectAllowedToolNamesParams["clientTools"];
|
||||
explicitAllowlistSources: Array<{ entries: string[] }>;
|
||||
}): string[] {
|
||||
const explicitNames = new Set(
|
||||
params.explicitAllowlistSources.flatMap((source) =>
|
||||
source.entries.map((entry) => normalizeToolName(entry)),
|
||||
),
|
||||
);
|
||||
return (params.clientTools ?? [])
|
||||
.map((tool) => tool.function?.name)
|
||||
.filter((name): name is string => Boolean(name?.trim()))
|
||||
.filter((name) => explicitNames.has(normalizeToolName(name)));
|
||||
}
|
||||
|
||||
export function buildToolSearchRunPlan(params: {
|
||||
visibleTools: CollectAllowedToolNamesParams["tools"];
|
||||
uncompactedTools: CollectAllowedToolNamesParams["tools"];
|
||||
clientTools?: CollectAllowedToolNamesParams["clientTools"];
|
||||
catalogRegistered: boolean;
|
||||
catalogToolCount: number;
|
||||
controlsEnabled: boolean;
|
||||
explicitAllowlistSources: Array<{ entries: string[] }>;
|
||||
}): ToolSearchRunPlan {
|
||||
const visibleAllowedToolNames = collectAllowedToolNames({
|
||||
tools: params.visibleTools,
|
||||
clientTools: params.catalogRegistered ? undefined : params.clientTools,
|
||||
});
|
||||
const replayAllowedToolNames = collectAllowedToolNames({
|
||||
tools: params.uncompactedTools,
|
||||
clientTools: params.clientTools,
|
||||
});
|
||||
const autoAddedControlNames = buildAutoAddedToolSearchControlNamesForAllowlistCheck({
|
||||
toolSearchControlsEnabled: params.controlsEnabled,
|
||||
explicitAllowlistSources: params.explicitAllowlistSources,
|
||||
});
|
||||
const clientCatalogCallableNames = params.catalogRegistered
|
||||
? collectExplicitlyAllowedClientToolNames({
|
||||
clientTools: params.clientTools,
|
||||
explicitAllowlistSources: params.explicitAllowlistSources,
|
||||
}).map((name) => `tool-search-client:${name}`)
|
||||
: [];
|
||||
return {
|
||||
visibleAllowedToolNames,
|
||||
replayAllowedToolNames,
|
||||
autoAddedControlNames,
|
||||
emptyAllowlistCallableNames: [
|
||||
...buildCallableToolNamesForEmptyAllowlistCheck({
|
||||
effectiveToolNames: [...visibleAllowedToolNames],
|
||||
autoAddedToolSearchControlNames: autoAddedControlNames,
|
||||
toolSearchCatalogToolCount: params.catalogToolCount,
|
||||
}),
|
||||
...clientCatalogCallableNames,
|
||||
],
|
||||
};
|
||||
}
|
||||
|
||||
export function resolveUnknownToolGuardThreshold(loopDetection?: {
|
||||
enabled?: boolean;
|
||||
@@ -962,25 +837,7 @@ export async function runEmbeddedAttempt(
|
||||
isRawModelRun,
|
||||
toolsAllow: params.toolsAllow,
|
||||
});
|
||||
const toolsEnabled = supportsModelTools(params.model);
|
||||
const toolSearchControlsEnabledForRun =
|
||||
toolsEnabled &&
|
||||
params.disableTools !== true &&
|
||||
!isRawModelRun &&
|
||||
params.toolsAllow?.length !== 0 &&
|
||||
resolveToolSearchConfig(params.config).enabled;
|
||||
const effectiveToolsAllow =
|
||||
toolSearchControlsEnabledForRun && params.toolsAllow
|
||||
? [...new Set([...params.toolsAllow, ...TOOL_SEARCH_CONTROL_ALLOWLIST_NAMES])]
|
||||
: params.toolsAllow;
|
||||
const shouldConstructTools =
|
||||
toolConstructionPlan.constructTools || toolSearchControlsEnabledForRun;
|
||||
let toolSearchCatalogExecutor: ToolSearchCatalogToolExecutor | undefined;
|
||||
const toolSearchCatalogRef: ToolSearchCatalogRef | undefined = toolSearchControlsEnabledForRun
|
||||
? createToolSearchCatalogRef()
|
||||
: undefined;
|
||||
const toolSearchTargetTranscriptProjections: ToolSearchTargetTranscriptProjection[] = [];
|
||||
const toolsRaw = !shouldConstructTools
|
||||
const toolsRaw = !toolConstructionPlan.constructTools
|
||||
? []
|
||||
: (() => {
|
||||
const allTools = createOpenClawCodingTools({
|
||||
@@ -1017,7 +874,6 @@ export async function runEmbeddedAttempt(
|
||||
: undefined,
|
||||
sessionId: params.sessionId,
|
||||
runId: params.runId,
|
||||
toolSearchCatalogRef,
|
||||
agentDir,
|
||||
workspaceDir: effectiveWorkspace,
|
||||
// When sandboxing uses a copied workspace (`ro` or `none`), effectiveWorkspace points
|
||||
@@ -1040,13 +896,6 @@ export async function runEmbeddedAttempt(
|
||||
currentThreadTs: params.currentThreadTs,
|
||||
currentMessageId: params.currentMessageId,
|
||||
includeCoreTools: toolConstructionPlan.includeCoreTools,
|
||||
includeToolSearchControls: true,
|
||||
toolSearchCatalogExecutor: (toolParams) => {
|
||||
if (!toolSearchCatalogExecutor) {
|
||||
throw new Error("Tool Search catalog executor is unavailable for this run.");
|
||||
}
|
||||
return toolSearchCatalogExecutor(toolParams);
|
||||
},
|
||||
toolConstructionPlan: toolConstructionPlan.codingToolConstructionPlan,
|
||||
replyToMode: params.replyToMode,
|
||||
hasRepliedRef: params.hasRepliedRef,
|
||||
@@ -1070,7 +919,7 @@ export async function runEmbeddedAttempt(
|
||||
},
|
||||
});
|
||||
corePluginToolStages.mark("attempt:create-openclaw-coding-tools");
|
||||
const filteredTools = applyEmbeddedAttemptToolsAllow(allTools, effectiveToolsAllow, {
|
||||
const filteredTools = applyEmbeddedAttemptToolsAllow(allTools, params.toolsAllow, {
|
||||
toolMeta: (tool) => getPluginToolMeta(tool),
|
||||
});
|
||||
corePluginToolStages.mark("attempt:tools-allow");
|
||||
@@ -1078,6 +927,7 @@ export async function runEmbeddedAttempt(
|
||||
})();
|
||||
prepStages.mark("core-plugin-tools");
|
||||
emitCorePluginToolStageSummary("core-plugin-tools", corePluginToolStages.snapshot());
|
||||
const toolsEnabled = supportsModelTools(params.model);
|
||||
const bootstrapHasFileAccess = toolsEnabled && toolsRaw.some((tool) => tool.name === "read");
|
||||
const bootstrapWarn = makeBootstrapWarn({
|
||||
sessionLabel,
|
||||
@@ -1275,40 +1125,12 @@ export async function runEmbeddedAttempt(
|
||||
ownerOnlyToolAllowlist: params.ownerOnlyToolAllowlist,
|
||||
warn: (message) => log.warn(message),
|
||||
});
|
||||
const uncompactedEffectiveTools = [...tools, ...filteredBundledTools];
|
||||
let effectiveTools = uncompactedEffectiveTools;
|
||||
const toolSearch = applyToolSearchCatalog({
|
||||
tools: effectiveTools,
|
||||
config: params.config,
|
||||
sessionId: params.sessionId,
|
||||
sessionKey: sandboxSessionKey,
|
||||
agentId: sessionAgentId,
|
||||
runId: params.runId,
|
||||
catalogRef: toolSearchCatalogRef,
|
||||
toolHookContext: {
|
||||
agentId: sessionAgentId,
|
||||
config: params.config,
|
||||
cwd: effectiveWorkspace,
|
||||
sessionKey: sandboxSessionKey,
|
||||
sessionId: params.sessionId,
|
||||
runId: params.runId,
|
||||
channelId: params.currentChannelId,
|
||||
trace: runTrace,
|
||||
loopDetection: resolveToolLoopDetectionConfig({
|
||||
cfg: params.config,
|
||||
agentId: sessionAgentId,
|
||||
}),
|
||||
onToolOutcome: params.onToolOutcome,
|
||||
},
|
||||
});
|
||||
effectiveTools = toolSearch.tools;
|
||||
if (toolSearch.compacted) {
|
||||
prepStages.mark("tool-search");
|
||||
log.info(
|
||||
`tool-search: cataloged ${toolSearch.catalogToolCount} tools behind compact prompt surface`,
|
||||
);
|
||||
}
|
||||
const effectiveTools = [...tools, ...filteredBundledTools];
|
||||
prepStages.mark("bundle-tools");
|
||||
const allowedToolNames = collectAllowedToolNames({
|
||||
tools: effectiveTools,
|
||||
clientTools,
|
||||
});
|
||||
const explicitToolAllowlistSources = collectAttemptExplicitToolAllowlistSources({
|
||||
config: params.config,
|
||||
sessionKey: params.sessionKey,
|
||||
@@ -1329,20 +1151,9 @@ export async function runEmbeddedAttempt(
|
||||
sandboxToolPolicy: sandbox?.tools,
|
||||
toolsAllow: params.toolsAllow,
|
||||
});
|
||||
const toolSearchRunPlan = buildToolSearchRunPlan({
|
||||
visibleTools: effectiveTools,
|
||||
uncompactedTools: uncompactedEffectiveTools,
|
||||
clientTools,
|
||||
catalogRegistered: toolSearch.catalogRegistered,
|
||||
catalogToolCount: toolSearch.catalogToolCount,
|
||||
controlsEnabled: toolSearchControlsEnabledForRun,
|
||||
explicitAllowlistSources: explicitToolAllowlistSources,
|
||||
});
|
||||
const allowedToolNames = toolSearchRunPlan.visibleAllowedToolNames;
|
||||
const replayAllowedToolNames = toolSearchRunPlan.replayAllowedToolNames;
|
||||
const emptyExplicitToolAllowlistError = buildEmptyExplicitToolAllowlistError({
|
||||
sources: explicitToolAllowlistSources,
|
||||
callableToolNames: toolSearchRunPlan.emptyAllowlistCallableNames,
|
||||
callableToolNames: effectiveTools.map((tool) => tool.name),
|
||||
toolsEnabled,
|
||||
disableTools: params.disableTools,
|
||||
});
|
||||
@@ -1636,7 +1447,7 @@ export async function runEmbeddedAttempt(
|
||||
params.model.api === "openai-codex-responses"
|
||||
? "aborted"
|
||||
: undefined,
|
||||
allowedToolNames: replayAllowedToolNames,
|
||||
allowedToolNames,
|
||||
suppressNextUserMessagePersistence: params.suppressNextUserMessagePersistence,
|
||||
onUserMessagePersisted: (message) => {
|
||||
params.onUserMessagePersisted?.(message);
|
||||
@@ -1772,7 +1583,7 @@ export async function runEmbeddedAttempt(
|
||||
// MEDIA: passthrough gate: a normalized alias is not sufficient — the
|
||||
// emitted tool name must match an exact registration of this run.
|
||||
const builtinToolNames = new Set(
|
||||
uncompactedEffectiveTools.flatMap((tool) => {
|
||||
effectiveTools.flatMap((tool) => {
|
||||
const name = (tool.name ?? "").trim();
|
||||
return name ? [name] : [];
|
||||
}),
|
||||
@@ -1782,10 +1593,15 @@ export async function runEmbeddedAttempt(
|
||||
// plugin tool names. MEDIA passthrough is still gated by the raw-name
|
||||
// set above, so a client tool that normalize-collides with a plugin
|
||||
// tool cannot inherit the plugin's local-media trust.
|
||||
const coreBuiltinToolNames = collectCoreBuiltinToolNames(uncompactedEffectiveTools, {
|
||||
isPluginTool: (tool) =>
|
||||
Boolean(getPluginToolMeta(tool as Parameters<typeof getPluginToolMeta>[0])),
|
||||
});
|
||||
const coreBuiltinToolNames = new Set(
|
||||
effectiveTools.flatMap((tool) => {
|
||||
const name = (tool.name ?? "").trim();
|
||||
if (!name || getPluginToolMeta(tool)) {
|
||||
return [];
|
||||
}
|
||||
return [name];
|
||||
}),
|
||||
);
|
||||
const clientToolNameConflicts = findClientToolNameConflicts({
|
||||
tools: clientTools ?? [],
|
||||
existingToolNames: [...coreBuiltinToolNames, ...PI_RESERVED_TOOL_NAMES],
|
||||
@@ -1793,7 +1609,7 @@ export async function runEmbeddedAttempt(
|
||||
if (clientToolNameConflicts.length > 0) {
|
||||
throw createClientToolNameConflictError(clientToolNameConflicts);
|
||||
}
|
||||
let clientToolDefs = clientTools
|
||||
const clientToolDefs = clientTools
|
||||
? toClientToolDefinitions(
|
||||
clientTools,
|
||||
{
|
||||
@@ -1835,21 +1651,6 @@ export async function runEmbeddedAttempt(
|
||||
},
|
||||
)
|
||||
: [];
|
||||
const clientToolSearch = addClientToolsToToolSearchCatalog({
|
||||
tools: clientToolDefs,
|
||||
config: params.config,
|
||||
sessionId: params.sessionId,
|
||||
sessionKey: sandboxSessionKey,
|
||||
agentId: sessionAgentId,
|
||||
runId: params.runId,
|
||||
catalogRef: toolSearchCatalogRef,
|
||||
});
|
||||
clientToolDefs = clientToolSearch.tools;
|
||||
if (clientToolSearch.compacted) {
|
||||
log.info(
|
||||
`tool-search: cataloged ${clientToolSearch.catalogToolCount} client tools behind compact prompt surface`,
|
||||
);
|
||||
}
|
||||
|
||||
const allCustomTools = [...customTools, ...clientToolDefs];
|
||||
// Pi treats `tools` as a name allowlist during session creation. Pass the
|
||||
@@ -2242,7 +2043,7 @@ export async function runEmbeddedAttempt(
|
||||
const nextMessages = sanitizeReplayToolCallIdsForStream({
|
||||
messages: messages as AgentMessage[],
|
||||
mode,
|
||||
allowedToolNames: replayAllowedToolNames,
|
||||
allowedToolNames,
|
||||
preserveNativeAnthropicToolUseIds: transcriptPolicy.preserveNativeAnthropicToolUseIds,
|
||||
preserveReplaySafeThinkingToolCallIds: shouldAllowProviderOwnedThinkingReplay({
|
||||
modelApi: (model as { api?: unknown })?.api as string | null | undefined,
|
||||
@@ -2399,7 +2200,7 @@ export async function runEmbeddedAttempt(
|
||||
modelApi: params.model.api,
|
||||
modelId: params.modelId,
|
||||
provider: params.provider,
|
||||
allowedToolNames: replayAllowedToolNames,
|
||||
allowedToolNames,
|
||||
config: params.config,
|
||||
workspaceDir: effectiveWorkspace,
|
||||
env: process.env,
|
||||
@@ -2646,7 +2447,6 @@ export async function runEmbeddedAttempt(
|
||||
const {
|
||||
assistantTexts,
|
||||
toolMetas,
|
||||
runToolLifecycle,
|
||||
unsubscribe,
|
||||
waitForCompactionRetry,
|
||||
isCompactionInFlight,
|
||||
@@ -2665,47 +2465,6 @@ export async function runEmbeddedAttempt(
|
||||
getCompactionCount,
|
||||
getLastCompactionTokensAfter,
|
||||
} = subscription;
|
||||
toolSearchCatalogExecutor = async (toolParams) => {
|
||||
try {
|
||||
const result = await runToolLifecycle({
|
||||
toolName: toolParams.toolName,
|
||||
toolCallId: toolParams.toolCallId,
|
||||
args: toolParams.input,
|
||||
execute: async () =>
|
||||
await toolParams.tool.execute(
|
||||
toolParams.toolCallId,
|
||||
toolParams.input,
|
||||
toolParams.signal ?? runAbortController.signal,
|
||||
toolParams.onUpdate,
|
||||
undefined as never,
|
||||
),
|
||||
});
|
||||
toolSearchTargetTranscriptProjections.push({
|
||||
parentToolCallId: toolParams.parentToolCallId,
|
||||
toolCallId: toolParams.toolCallId,
|
||||
toolName: toolParams.toolName,
|
||||
input: toolParams.input,
|
||||
result,
|
||||
timestamp: Date.now(),
|
||||
});
|
||||
return result;
|
||||
} catch (error) {
|
||||
const message = formatErrorMessage(error);
|
||||
toolSearchTargetTranscriptProjections.push({
|
||||
parentToolCallId: toolParams.parentToolCallId,
|
||||
toolCallId: toolParams.toolCallId,
|
||||
toolName: toolParams.toolName,
|
||||
input: toolParams.input,
|
||||
result: {
|
||||
content: [{ type: "text", text: message }],
|
||||
details: { status: "error", error: message },
|
||||
},
|
||||
isError: true,
|
||||
timestamp: Date.now(),
|
||||
});
|
||||
throw error;
|
||||
}
|
||||
};
|
||||
|
||||
const queueHandle: EmbeddedPiQueueHandle & {
|
||||
kind: "embedded";
|
||||
@@ -3260,17 +3019,11 @@ export async function runEmbeddedAttempt(
|
||||
messages: activeSession.messages,
|
||||
note: `images: prompt=${imageResult.images.length}`,
|
||||
});
|
||||
const trajectoryProviderVisibleTools = toTrajectoryToolDefinitions(effectiveTools);
|
||||
trajectoryRecorder?.recordEvent("context.compiled", {
|
||||
systemPrompt: systemPromptForHook,
|
||||
prompt: promptForModel,
|
||||
messages: activeSession.messages,
|
||||
tools: toTrajectoryToolDefinitions(
|
||||
toolSearch.compacted ? uncompactedEffectiveTools : effectiveTools,
|
||||
),
|
||||
...(toolSearch.compacted
|
||||
? { providerVisibleTools: trajectoryProviderVisibleTools }
|
||||
: {}),
|
||||
tools: toTrajectoryToolDefinitions(effectiveTools),
|
||||
imagesCount: imageResult.images.length,
|
||||
streamStrategy,
|
||||
transport: effectiveAgentTransport,
|
||||
@@ -3644,10 +3397,7 @@ export async function runEmbeddedAttempt(
|
||||
);
|
||||
}
|
||||
}
|
||||
messagesSnapshot = projectToolSearchTargetTranscriptMessages(
|
||||
snapshotSelection.messagesSnapshot,
|
||||
toolSearchTargetTranscriptProjections,
|
||||
);
|
||||
messagesSnapshot = snapshotSelection.messagesSnapshot;
|
||||
sessionIdUsed = snapshotSelection.sessionIdUsed;
|
||||
|
||||
lastAssistant = messagesSnapshot
|
||||
@@ -4109,13 +3859,6 @@ export async function runEmbeddedAttempt(
|
||||
// See: https://github.com/openclaw/openclaw/issues/8643
|
||||
let cleanupError: unknown;
|
||||
try {
|
||||
clearToolSearchCatalog({
|
||||
sessionId: params.sessionId,
|
||||
sessionKey: sandboxSessionKey,
|
||||
agentId: sessionAgentId,
|
||||
runId: params.runId,
|
||||
catalogRef: toolSearchCatalogRef,
|
||||
});
|
||||
await cleanupEmbeddedAttemptResources({
|
||||
removeToolResultContextGuard,
|
||||
flushPendingToolResultsAfterIdle,
|
||||
|
||||
@@ -119,27 +119,6 @@ describe("resolveLlmIdleTimeoutMs", () => {
|
||||
expect(resolveLlmIdleTimeoutMs({ model: { baseUrl } })).toBe(0);
|
||||
});
|
||||
|
||||
it("keeps the default idle watchdog for Ollama cloud models routed through local Ollama", () => {
|
||||
expect(
|
||||
resolveLlmIdleTimeoutMs({
|
||||
model: {
|
||||
provider: "ollama",
|
||||
id: "glm-5.1:cloud",
|
||||
baseUrl: "http://127.0.0.1:11434",
|
||||
},
|
||||
}),
|
||||
).toBe(DEFAULT_LLM_IDLE_TIMEOUT_MS);
|
||||
expect(
|
||||
resolveLlmIdleTimeoutMs({
|
||||
model: {
|
||||
provider: "ollama2",
|
||||
id: "ollama2/kimi-k2.5:cloud",
|
||||
baseUrl: "http://localhost:11434",
|
||||
},
|
||||
}),
|
||||
).toBe(DEFAULT_LLM_IDLE_TIMEOUT_MS);
|
||||
});
|
||||
|
||||
it.each([
|
||||
"http://172.32.0.1:11434",
|
||||
"http://192.169.1.1:11434",
|
||||
|
||||
@@ -91,23 +91,6 @@ function isLocalProviderBaseUrl(baseUrl: string): boolean {
|
||||
);
|
||||
}
|
||||
|
||||
function isOllamaCloudModel(model: { id?: string; provider?: string } | undefined): boolean {
|
||||
const rawModelId = model?.id;
|
||||
if (typeof rawModelId !== "string") {
|
||||
return false;
|
||||
}
|
||||
|
||||
const provider = model?.provider?.trim().toLowerCase();
|
||||
if (provider && !provider.startsWith("ollama")) {
|
||||
return false;
|
||||
}
|
||||
|
||||
const modelId = rawModelId.trim().toLowerCase();
|
||||
const slashIndex = modelId.indexOf("/");
|
||||
const bareModelId = slashIndex >= 0 ? modelId.slice(slashIndex + 1) : modelId;
|
||||
return bareModelId.endsWith(":cloud");
|
||||
}
|
||||
|
||||
/**
|
||||
* Resolves the LLM idle timeout from configuration.
|
||||
* @returns Idle timeout in milliseconds, or 0 to disable
|
||||
@@ -117,7 +100,7 @@ export function resolveLlmIdleTimeoutMs(params?: {
|
||||
trigger?: EmbeddedRunTrigger;
|
||||
runTimeoutMs?: number;
|
||||
modelRequestTimeoutMs?: number;
|
||||
model?: { baseUrl?: string; id?: string; provider?: string };
|
||||
model?: { baseUrl?: string };
|
||||
}): number {
|
||||
const clampTimeoutMs = (valueMs: number) => Math.min(Math.floor(valueMs), MAX_SAFE_TIMEOUT_MS);
|
||||
const clampImplicitTimeoutMs = (valueMs: number) =>
|
||||
@@ -173,16 +156,9 @@ export function resolveLlmIdleTimeoutMs(params?: {
|
||||
// Local providers can legitimately stream nothing for many minutes during
|
||||
// prompt evaluation or thinking, so falling back to the default would abort
|
||||
// valid local runs. Honor it only when the user has not opted out via the
|
||||
// baseUrl pointing at loopback / private-network / `.local`. Ollama cloud
|
||||
// models are still hosted remotely even when proxied through local Ollama, so
|
||||
// keep the cloud watchdog for `*:cloud` model ids.
|
||||
// baseUrl pointing at loopback / private-network / `.local`.
|
||||
const baseUrl = params?.model?.baseUrl;
|
||||
if (
|
||||
typeof baseUrl === "string" &&
|
||||
baseUrl.length > 0 &&
|
||||
isLocalProviderBaseUrl(baseUrl) &&
|
||||
!isOllamaCloudModel(params?.model)
|
||||
) {
|
||||
if (typeof baseUrl === "string" && baseUrl.length > 0 && isLocalProviderBaseUrl(baseUrl)) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
|
||||
@@ -9,10 +9,7 @@ import type { ReplyOperation } from "../../../auto-reply/reply/reply-run-registr
|
||||
import type { ReasoningLevel, ThinkLevel, VerboseLevel } from "../../../auto-reply/thinking.js";
|
||||
import type { OpenClawConfig } from "../../../config/types.openclaw.js";
|
||||
import type { PromptImageOrderEntry } from "../../../media/prompt-image-order.js";
|
||||
import type {
|
||||
CommandPriority,
|
||||
CommandQueueEnqueueFn,
|
||||
} from "../../../process/command-queue.types.js";
|
||||
import type { CommandQueueEnqueueFn } from "../../../process/command-queue.types.js";
|
||||
import type { InputProvenance } from "../../../sessions/input-provenance.js";
|
||||
import type { ExecElevatedDefaults, ExecToolDefaults } from "../../bash-tools.exec-types.js";
|
||||
import type { AgentStreamParams, ClientToolDefinition } from "../../command/shared-types.js";
|
||||
@@ -187,7 +184,6 @@ export type RunEmbeddedPiAgentParams = {
|
||||
sessionKey?: string;
|
||||
}) => void | Promise<void>;
|
||||
lane?: string;
|
||||
queuePriority?: CommandPriority;
|
||||
enqueue?: CommandQueueEnqueueFn;
|
||||
extraSystemPrompt?: string;
|
||||
sourceReplyDeliveryMode?: SourceReplyDeliveryMode;
|
||||
|
||||
@@ -185,8 +185,7 @@ describe("buildEmbeddedSystemPrompt", () => {
|
||||
|
||||
expect(prompt).toContain("Active background exec sessions in this scope:");
|
||||
expect(prompt).toContain("sess-active running pid=1234 cwd=/tmp/work :: sleep 600");
|
||||
expect(prompt).toContain("Use process log before interactive input");
|
||||
expect(prompt).toContain("waitingForInput/stdinWritable");
|
||||
expect(prompt).toContain("process tool with a sessionId");
|
||||
expect(prompt).toContain("process list");
|
||||
});
|
||||
});
|
||||
|
||||
@@ -1,15 +1,7 @@
|
||||
import { describe, expect, it } from "vitest";
|
||||
import { findClientToolNameConflicts } from "../pi-tool-definition-adapter.js";
|
||||
import { createStubTool } from "../test-helpers/pi-tool-stubs.js";
|
||||
import {
|
||||
addClientToolsToToolSearchCatalog,
|
||||
applyToolSearchCatalog,
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
} from "../tool-search.js";
|
||||
import type { ClientToolDefinition } from "./run/params.js";
|
||||
import {
|
||||
collectAllowedToolNames,
|
||||
collectCoreBuiltinToolNames,
|
||||
collectRegisteredToolNames,
|
||||
PI_RESERVED_TOOL_NAMES,
|
||||
toSessionToolAllowlist,
|
||||
@@ -53,37 +45,6 @@ describe("tool name allowlists", () => {
|
||||
expect(allowlist).toEqual(["exec", "image_generate", "read"]);
|
||||
});
|
||||
|
||||
it("keeps hidden core names available for client conflict admission", () => {
|
||||
const uncompactedTools = [
|
||||
createStubTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME),
|
||||
createStubTool("exec"),
|
||||
createStubTool("message"),
|
||||
];
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: uncompactedTools,
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-conflict-admission",
|
||||
});
|
||||
const names = collectCoreBuiltinToolNames(uncompactedTools);
|
||||
|
||||
expect([...names]).toEqual([TOOL_SEARCH_CODE_MODE_TOOL_NAME, "exec", "message"]);
|
||||
expect(compacted.tools.map((tool) => tool.name)).toEqual([TOOL_SEARCH_CODE_MODE_TOOL_NAME]);
|
||||
expect(
|
||||
findClientToolNameConflicts({
|
||||
tools: [
|
||||
{
|
||||
type: "function",
|
||||
function: {
|
||||
name: "exec",
|
||||
parameters: { type: "object", properties: {} },
|
||||
},
|
||||
},
|
||||
],
|
||||
existingToolNames: [...names, ...PI_RESERVED_TOOL_NAMES],
|
||||
}),
|
||||
).toEqual(["exec"]);
|
||||
});
|
||||
|
||||
it("pins the reserved Pi built-in tool namespace used by client conflict checks", () => {
|
||||
expect(PI_RESERVED_TOOL_NAMES).toEqual(["bash", "edit", "find", "grep", "ls", "read", "write"]);
|
||||
});
|
||||
@@ -106,82 +67,4 @@ describe("tool name allowlists", () => {
|
||||
|
||||
expect(allowlist).toEqual(["exec", "image_generate", "read"]);
|
||||
});
|
||||
|
||||
it("excludes client tool names when Tool Search compacts them into the catalog", () => {
|
||||
const config = { tools: { toolSearch: true } } as never;
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: [createStubTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME)],
|
||||
config,
|
||||
sessionId: "session-client-allowed-names",
|
||||
});
|
||||
const clientTools: ClientToolDefinition[] = [
|
||||
{
|
||||
type: "function",
|
||||
function: {
|
||||
name: "client_pick_file",
|
||||
parameters: { type: "object", properties: {} },
|
||||
},
|
||||
},
|
||||
];
|
||||
const clientToolSearch = addClientToolsToToolSearchCatalog({
|
||||
tools: [createStubTool("client_pick_file")],
|
||||
config,
|
||||
sessionId: "session-client-allowed-names",
|
||||
});
|
||||
|
||||
const allowlist = toSessionToolAllowlist(
|
||||
collectAllowedToolNames({
|
||||
tools: compacted.tools,
|
||||
clientTools: compacted.catalogRegistered ? undefined : clientTools,
|
||||
}),
|
||||
);
|
||||
|
||||
expect(compacted.catalogRegistered).toBe(true);
|
||||
expect(clientToolSearch.tools).toEqual([]);
|
||||
expect(allowlist).toEqual([TOOL_SEARCH_CODE_MODE_TOOL_NAME]);
|
||||
});
|
||||
|
||||
it("keeps hidden catalog tools valid for replay guards after Tool Search compaction", () => {
|
||||
const config = { tools: { toolSearch: true } } as never;
|
||||
const uncompactedTools = [
|
||||
createStubTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME),
|
||||
createStubTool("exec"),
|
||||
createStubTool("fake_plugin_tool"),
|
||||
];
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: uncompactedTools,
|
||||
config,
|
||||
sessionId: "session-replay-allowed-names",
|
||||
});
|
||||
const clientTools: ClientToolDefinition[] = [
|
||||
{
|
||||
type: "function",
|
||||
function: {
|
||||
name: "client_pick_file",
|
||||
parameters: { type: "object", properties: {} },
|
||||
},
|
||||
},
|
||||
];
|
||||
|
||||
const visibleAllowlist = toSessionToolAllowlist(
|
||||
collectAllowedToolNames({
|
||||
tools: compacted.tools,
|
||||
clientTools: compacted.catalogRegistered ? undefined : clientTools,
|
||||
}),
|
||||
);
|
||||
const replayAllowlist = toSessionToolAllowlist(
|
||||
collectAllowedToolNames({
|
||||
tools: uncompactedTools,
|
||||
clientTools,
|
||||
}),
|
||||
);
|
||||
|
||||
expect(visibleAllowlist).toEqual([TOOL_SEARCH_CODE_MODE_TOOL_NAME]);
|
||||
expect(replayAllowlist).toEqual([
|
||||
"client_pick_file",
|
||||
"exec",
|
||||
"fake_plugin_tool",
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
]);
|
||||
});
|
||||
});
|
||||
|
||||
@@ -42,20 +42,6 @@ export function collectRegisteredToolNames(tools: Array<{ name?: string }>): Set
|
||||
return names;
|
||||
}
|
||||
|
||||
export function collectCoreBuiltinToolNames(
|
||||
tools: Array<{ name?: string }>,
|
||||
options?: { isPluginTool?: (tool: { name?: string }) => boolean },
|
||||
): Set<string> {
|
||||
const names = new Set<string>();
|
||||
for (const tool of tools) {
|
||||
if (options?.isPluginTool?.(tool)) {
|
||||
continue;
|
||||
}
|
||||
addName(names, tool.name);
|
||||
}
|
||||
return names;
|
||||
}
|
||||
|
||||
export function toSessionToolAllowlist(allowedToolNames: Iterable<string>): string[] {
|
||||
return [...new Set(allowedToolNames)].toSorted((a, b) => a.localeCompare(b));
|
||||
}
|
||||
|
||||
@@ -28,10 +28,6 @@ import {
|
||||
consumePendingToolMediaIntoReply,
|
||||
readPendingToolMediaReply,
|
||||
} from "./pi-embedded-subscribe.handlers.messages.js";
|
||||
import {
|
||||
handleToolExecutionEnd,
|
||||
handleToolExecutionStart,
|
||||
} from "./pi-embedded-subscribe.handlers.tools.js";
|
||||
import type {
|
||||
EmbeddedPiSubscribeContext,
|
||||
EmbeddedPiSubscribeState,
|
||||
@@ -1028,44 +1024,6 @@ export function subscribeEmbeddedPiSession(params: SubscribeEmbeddedPiSessionPar
|
||||
return {
|
||||
assistantTexts,
|
||||
toolMetas,
|
||||
runToolLifecycle: async <T>(toolParams: {
|
||||
toolName: string;
|
||||
toolCallId: string;
|
||||
args: unknown;
|
||||
execute: () => Promise<T>;
|
||||
}): Promise<T> => {
|
||||
await handleToolExecutionStart(ctx, {
|
||||
type: "tool_execution_start",
|
||||
toolName: toolParams.toolName,
|
||||
toolCallId: toolParams.toolCallId,
|
||||
args: toolParams.args,
|
||||
} as never);
|
||||
try {
|
||||
const result = await toolParams.execute();
|
||||
await handleToolExecutionEnd(ctx, {
|
||||
type: "tool_execution_end",
|
||||
toolName: toolParams.toolName,
|
||||
toolCallId: toolParams.toolCallId,
|
||||
isError: false,
|
||||
result,
|
||||
} as never);
|
||||
return result;
|
||||
} catch (error) {
|
||||
await handleToolExecutionEnd(ctx, {
|
||||
type: "tool_execution_end",
|
||||
toolName: toolParams.toolName,
|
||||
toolCallId: toolParams.toolCallId,
|
||||
isError: true,
|
||||
result: {
|
||||
details: {
|
||||
status: "error",
|
||||
error: error instanceof Error ? error.message : String(error),
|
||||
},
|
||||
},
|
||||
} as never);
|
||||
throw error;
|
||||
}
|
||||
},
|
||||
unsubscribe,
|
||||
setTerminalLifecycleMeta: (meta: {
|
||||
replayInvalid?: boolean;
|
||||
|
||||
@@ -1,7 +1,6 @@
|
||||
import { copyPluginToolMeta } from "../plugins/tools.js";
|
||||
import { bindAbortRelay } from "../utils/fetch-timeout.js";
|
||||
import { copyChannelAgentToolMeta } from "./channel-tools.js";
|
||||
import { copyBeforeToolCallHookMarker } from "./pi-tools.before-tool-call.js";
|
||||
import type { AnyAgentTool } from "./pi-tools.types.js";
|
||||
|
||||
function throwAbortError(): never {
|
||||
@@ -69,6 +68,5 @@ export function wrapToolWithAbortSignal(
|
||||
};
|
||||
copyPluginToolMeta(tool, wrappedTool);
|
||||
copyChannelAgentToolMeta(tool as never, wrappedTool as never);
|
||||
copyBeforeToolCallHookMarker(tool, wrappedTool);
|
||||
return wrappedTool;
|
||||
}
|
||||
|
||||
@@ -18,7 +18,6 @@ import { wrapToolWithAbortSignal } from "./pi-tools.abort.js";
|
||||
import {
|
||||
__testing as beforeToolCallTesting,
|
||||
consumeAdjustedParamsForToolCall,
|
||||
isToolWrappedWithBeforeToolCallHook,
|
||||
wrapToolWithBeforeToolCallHook,
|
||||
} from "./pi-tools.before-tool-call.js";
|
||||
|
||||
@@ -312,18 +311,6 @@ describe("before_tool_call hook deduplication (#15502)", () => {
|
||||
|
||||
expect(beforeToolCallHook).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
|
||||
it("preserves the hook marker when abort wrapping a hooked tool", () => {
|
||||
const execute = vi.fn().mockResolvedValue({ content: [], details: { ok: true } });
|
||||
const baseTool = { name: "Bash", execute, description: "bash", parameters: {} } as any;
|
||||
const wrapped = wrapToolWithBeforeToolCallHook(baseTool, {
|
||||
agentId: "main",
|
||||
sessionKey: "main",
|
||||
});
|
||||
const withAbort = wrapToolWithAbortSignal(wrapped, new AbortController().signal);
|
||||
|
||||
expect(isToolWrappedWithBeforeToolCallHook(withAbort)).toBe(true);
|
||||
});
|
||||
});
|
||||
|
||||
describe("before_tool_call hook integration for client tools", () => {
|
||||
|
||||
@@ -791,16 +791,6 @@ export function isToolWrappedWithBeforeToolCallHook(tool: AnyAgentTool): boolean
|
||||
return taggedTool[BEFORE_TOOL_CALL_WRAPPED] === true;
|
||||
}
|
||||
|
||||
export function copyBeforeToolCallHookMarker(source: AnyAgentTool, target: AnyAgentTool): void {
|
||||
if (!isToolWrappedWithBeforeToolCallHook(source)) {
|
||||
return;
|
||||
}
|
||||
Object.defineProperty(target, BEFORE_TOOL_CALL_WRAPPED, {
|
||||
value: true,
|
||||
enumerable: true,
|
||||
});
|
||||
}
|
||||
|
||||
export function consumeAdjustedParamsForToolCall(toolCallId: string, runId?: string): unknown {
|
||||
const adjustedParamsKey = buildAdjustedParamsKey({ runId, toolCallId });
|
||||
const params = adjustedParamsByToolCallId.get(adjustedParamsKey);
|
||||
|
||||
@@ -155,121 +155,6 @@ describe("createOpenClawCodingTools", () => {
|
||||
expectListIncludes([...values], ["restart", "config.get", "config.patch", "config.apply"]);
|
||||
});
|
||||
|
||||
it("does not add Tool Search control tools from the shared factory by default", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
config: {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
},
|
||||
});
|
||||
const names = new Set(tools.map((tool) => tool.name));
|
||||
|
||||
expect(names.has("tool_search_code")).toBe(false);
|
||||
expect(names.has("tool_search")).toBe(false);
|
||||
expect(names.has("tool_describe")).toBe(false);
|
||||
expect(names.has("tool_call")).toBe(false);
|
||||
});
|
||||
|
||||
it("adds PI Tool Search control tools when explicitly requested", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
includeToolSearchControls: true,
|
||||
config: {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
},
|
||||
});
|
||||
const names = new Set(tools.map((tool) => tool.name));
|
||||
|
||||
expect(names.has("tool_search_code")).toBe(true);
|
||||
expect(names.has("tool_search")).toBe(true);
|
||||
expect(names.has("tool_describe")).toBe(true);
|
||||
expect(names.has("tool_call")).toBe(true);
|
||||
});
|
||||
|
||||
it("keeps PI Tool Search controls available under restrictive tool profiles", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
includeToolSearchControls: true,
|
||||
config: {
|
||||
tools: {
|
||||
profile: "coding",
|
||||
toolSearch: true,
|
||||
},
|
||||
},
|
||||
});
|
||||
const names = new Set(tools.map((tool) => tool.name));
|
||||
|
||||
expect(names.has("tool_search_code")).toBe(true);
|
||||
expect(names.has("tool_search")).toBe(true);
|
||||
expect(names.has("tool_describe")).toBe(true);
|
||||
expect(names.has("tool_call")).toBe(true);
|
||||
expect(names.has("message")).toBe(false);
|
||||
});
|
||||
|
||||
it("keeps PI Tool Search controls available under restrictive tool allowlists", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
includeToolSearchControls: true,
|
||||
config: {
|
||||
tools: {
|
||||
allow: ["read"],
|
||||
toolSearch: true,
|
||||
},
|
||||
},
|
||||
});
|
||||
const names = new Set(tools.map((tool) => tool.name));
|
||||
|
||||
expect(names.has("read")).toBe(true);
|
||||
expect(names.has("exec")).toBe(false);
|
||||
expect(names.has("tool_search_code")).toBe(true);
|
||||
expect(names.has("tool_search")).toBe(true);
|
||||
expect(names.has("tool_describe")).toBe(true);
|
||||
expect(names.has("tool_call")).toBe(true);
|
||||
});
|
||||
|
||||
it("lets explicit deny policies remove PI Tool Search controls", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
includeToolSearchControls: true,
|
||||
config: {
|
||||
tools: {
|
||||
profile: "coding",
|
||||
deny: ["tool_search_code"],
|
||||
toolSearch: true,
|
||||
},
|
||||
},
|
||||
});
|
||||
const names = new Set(tools.map((tool) => tool.name));
|
||||
|
||||
expect(names.has("tool_search_code")).toBe(false);
|
||||
expect(names.has("read")).toBe(true);
|
||||
});
|
||||
|
||||
it("keeps PI Tool Search controls when core OpenClaw tools are not materialized", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
includeCoreTools: false,
|
||||
includeToolSearchControls: true,
|
||||
toolConstructionPlan: {
|
||||
includeBaseCodingTools: false,
|
||||
includeShellTools: false,
|
||||
includeChannelTools: false,
|
||||
includeOpenClawTools: false,
|
||||
includePluginTools: true,
|
||||
},
|
||||
config: {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
},
|
||||
});
|
||||
const names = new Set(tools.map((tool) => tool.name));
|
||||
|
||||
expect(names.has("tool_search_code")).toBe(true);
|
||||
expect(names.has("tool_search")).toBe(true);
|
||||
expect(names.has("tool_describe")).toBe(true);
|
||||
expect(names.has("tool_call")).toBe(true);
|
||||
expect(names.has("message")).toBe(false);
|
||||
});
|
||||
|
||||
it("exposes only an explicitly authorized owner-only tool to non-owner sessions", () => {
|
||||
const tools = createOpenClawCodingTools({
|
||||
config: testConfig,
|
||||
|
||||
@@ -78,16 +78,6 @@ import {
|
||||
normalizeToolName,
|
||||
resolveToolProfilePolicy,
|
||||
} from "./tool-policy.js";
|
||||
import {
|
||||
createToolSearchTools,
|
||||
resolveToolSearchConfig,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
type ToolSearchCatalogRef,
|
||||
type ToolSearchCatalogToolExecutor,
|
||||
} from "./tool-search.js";
|
||||
import { resolveWorkspaceRoot } from "./workspace-dir.js";
|
||||
|
||||
function isOpenAIProvider(provider?: string) {
|
||||
@@ -386,12 +376,6 @@ export function createOpenClawCodingTools(options?: {
|
||||
forceHeartbeatTool?: boolean;
|
||||
/** If false, build plugin tools only while preserving the shared policy pipeline. */
|
||||
includeCoreTools?: boolean;
|
||||
/** PI-only: expose OpenClaw Tool Search controls for catalog compaction. */
|
||||
includeToolSearchControls?: boolean;
|
||||
/** Executes cataloged tools through the active PI run lifecycle. */
|
||||
toolSearchCatalogExecutor?: ToolSearchCatalogToolExecutor;
|
||||
/** Runtime-local Tool Search catalog ref shared with PI attempt compaction. */
|
||||
toolSearchCatalogRef?: ToolSearchCatalogRef;
|
||||
/** Limits which tool families are materialized before the shared policy pipeline runs. */
|
||||
toolConstructionPlan?: OpenClawCodingToolConstructionPlan;
|
||||
/** Whether the sender is an owner (required for owner-only tools). */
|
||||
@@ -466,24 +450,9 @@ export function createOpenClawCodingTools(options?: {
|
||||
(options?.trigger === "heartbeat" &&
|
||||
options?.config?.messages?.visibleReplies === "message_tool");
|
||||
const forceHeartbeatTool = options?.forceHeartbeatTool === true || enableHeartbeatTool;
|
||||
const toolSearchConfig = resolveToolSearchConfig(options?.config);
|
||||
const toolSearchControlsEnabled =
|
||||
options?.includeToolSearchControls === true && toolSearchConfig.enabled;
|
||||
const toolSearchControlAllowlist = toolSearchControlsEnabled
|
||||
? [
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
]
|
||||
: [];
|
||||
const mergeToolSearchControlAllowlist = <TPolicy extends { allow?: string[] }>(
|
||||
policy: TPolicy | undefined,
|
||||
) => mergeAlsoAllowPolicy(policy, toolSearchControlAllowlist);
|
||||
const runtimeProfileAlsoAllow = [
|
||||
...(options?.forceMessageTool ? ["message"] : []),
|
||||
...(forceHeartbeatTool ? [HEARTBEAT_RESPONSE_TOOL_NAME] : []),
|
||||
...toolSearchControlAllowlist,
|
||||
];
|
||||
const profilePolicyWithAlsoAllow = mergeAlsoAllowPolicy(profilePolicy, [
|
||||
...(profileAlsoAllow ?? []),
|
||||
@@ -514,26 +483,16 @@ export function createOpenClawCodingTools(options?: {
|
||||
store: subagentStore,
|
||||
})
|
||||
: undefined;
|
||||
const globalPolicyWithToolSearchControls = mergeToolSearchControlAllowlist(globalPolicy);
|
||||
const globalProviderPolicyWithToolSearchControls =
|
||||
mergeToolSearchControlAllowlist(globalProviderPolicy);
|
||||
const agentPolicyWithToolSearchControls = mergeToolSearchControlAllowlist(agentPolicy);
|
||||
const agentProviderPolicyWithToolSearchControls =
|
||||
mergeToolSearchControlAllowlist(agentProviderPolicy);
|
||||
const groupPolicyWithToolSearchControls = mergeToolSearchControlAllowlist(groupPolicy);
|
||||
const sandboxToolPolicyWithToolSearchControls =
|
||||
mergeToolSearchControlAllowlist(sandboxToolPolicy);
|
||||
const subagentPolicyWithToolSearchControls = mergeToolSearchControlAllowlist(subagentPolicy);
|
||||
const allowBackground = isToolAllowedByPolicies("process", [
|
||||
profilePolicyWithAlsoAllow,
|
||||
providerProfilePolicyWithAlsoAllow,
|
||||
globalPolicyWithToolSearchControls,
|
||||
globalProviderPolicyWithToolSearchControls,
|
||||
agentPolicyWithToolSearchControls,
|
||||
agentProviderPolicyWithToolSearchControls,
|
||||
groupPolicyWithToolSearchControls,
|
||||
sandboxToolPolicyWithToolSearchControls,
|
||||
subagentPolicyWithToolSearchControls,
|
||||
globalPolicy,
|
||||
globalProviderPolicy,
|
||||
agentPolicy,
|
||||
agentProviderPolicy,
|
||||
groupPolicy,
|
||||
sandboxToolPolicy,
|
||||
subagentPolicy,
|
||||
]);
|
||||
options?.recordToolPrepStage?.("tool-policy");
|
||||
const execConfig = resolveExecConfig({ cfg: options?.config, agentId });
|
||||
@@ -746,19 +705,6 @@ export function createOpenClawCodingTools(options?: {
|
||||
},
|
||||
resolvedConfig: options?.config,
|
||||
});
|
||||
const toolSearchTools = toolSearchControlsEnabled
|
||||
? createToolSearchTools({
|
||||
config: options?.config,
|
||||
runtimeConfig: options?.config,
|
||||
agentId,
|
||||
sessionKey: options?.sessionKey,
|
||||
sessionId: options?.sessionId,
|
||||
runId: options?.runId,
|
||||
catalogRef: options?.toolSearchCatalogRef,
|
||||
abortSignal: options?.abortSignal,
|
||||
executeTool: options?.toolSearchCatalogExecutor,
|
||||
})
|
||||
: [];
|
||||
const tools: AnyAgentTool[] = [
|
||||
...base,
|
||||
...(includeBaseCodingTools && sandboxRoot
|
||||
@@ -842,7 +788,6 @@ export function createOpenClawCodingTools(options?: {
|
||||
recordToolPrepStage: options?.recordToolPrepStage,
|
||||
})
|
||||
: pluginToolsOnly),
|
||||
...toolSearchTools,
|
||||
];
|
||||
options?.recordToolPrepStage?.("openclaw-tools");
|
||||
const toolsForMemoryFlush: AnyAgentTool[] = isMemoryFlushRun && memoryFlushWritePath ? [] : tools;
|
||||
@@ -902,15 +847,15 @@ export function createOpenClawCodingTools(options?: {
|
||||
providerProfilePolicy: providerProfilePolicyWithAlsoAllow,
|
||||
providerProfile,
|
||||
providerProfileUnavailableCoreWarningAllowlist: providerProfilePolicy?.allow,
|
||||
globalPolicy: globalPolicyWithToolSearchControls,
|
||||
globalProviderPolicy: globalProviderPolicyWithToolSearchControls,
|
||||
agentPolicy: agentPolicyWithToolSearchControls,
|
||||
agentProviderPolicy: agentProviderPolicyWithToolSearchControls,
|
||||
groupPolicy: groupPolicyWithToolSearchControls,
|
||||
globalPolicy,
|
||||
globalProviderPolicy,
|
||||
agentPolicy,
|
||||
agentProviderPolicy,
|
||||
groupPolicy,
|
||||
agentId,
|
||||
}),
|
||||
{ policy: sandboxToolPolicyWithToolSearchControls, label: "sandbox tools.allow" },
|
||||
{ policy: subagentPolicyWithToolSearchControls, label: "subagent tools.allow" },
|
||||
{ policy: sandboxToolPolicy, label: "sandbox tools.allow" },
|
||||
{ policy: subagentPolicy, label: "subagent tools.allow" },
|
||||
],
|
||||
});
|
||||
options?.recordToolPrepStage?.("authorization-policy");
|
||||
|
||||
@@ -1299,7 +1299,7 @@ function buildActiveProcessSessionReferenceLines(
|
||||
const cwd = session.cwd ? ` cwd=${sanitizeForPromptLiteral(session.cwd)}` : "";
|
||||
return `- ${session.sessionId} ${session.status}${pid}${cwd} :: ${sanitizeForPromptLiteral(session.name)}`;
|
||||
}),
|
||||
"Use process log before interactive input; log/poll may report waitingForInput/stdinWritable. If prior context lost a sessionId, run process list.",
|
||||
"Use the process tool with a sessionId to poll, log, write to, or terminate these sessions. If prior context lost a sessionId, run process list.",
|
||||
];
|
||||
}
|
||||
|
||||
|
||||
@@ -1,858 +0,0 @@
|
||||
import { describe, expect, it, vi } from "vitest";
|
||||
import { setPluginToolMeta } from "../plugins/tools.js";
|
||||
import { wrapToolWithAbortSignal } from "./pi-tools.abort.js";
|
||||
import {
|
||||
isToolWrappedWithBeforeToolCallHook,
|
||||
wrapToolWithBeforeToolCallHook,
|
||||
} from "./pi-tools.before-tool-call.js";
|
||||
import {
|
||||
__testing,
|
||||
addClientToolsToToolSearchCatalog,
|
||||
applyToolSearchCatalog,
|
||||
clearToolSearchCatalog,
|
||||
createToolSearchCatalogRef,
|
||||
createToolSearchTools,
|
||||
projectToolSearchTargetTranscriptMessages,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_SEARCH_CODE_MODE_TOOL_NAME,
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
} from "./tool-search.js";
|
||||
import { jsonResult, type AnyAgentTool } from "./tools/common.js";
|
||||
|
||||
function fakeTool(name: string, description: string): AnyAgentTool {
|
||||
return {
|
||||
name,
|
||||
label: name,
|
||||
description,
|
||||
parameters: {
|
||||
type: "object",
|
||||
properties: {
|
||||
value: { type: "string" },
|
||||
},
|
||||
},
|
||||
execute: vi.fn(async (_toolCallId, input) => jsonResult({ name, input })),
|
||||
};
|
||||
}
|
||||
|
||||
function pluginTool(name: string, description: string, pluginId = "fake-catalog"): AnyAgentTool {
|
||||
const tool = fakeTool(name, description);
|
||||
setPluginToolMeta(tool, {
|
||||
pluginId,
|
||||
optional: true,
|
||||
});
|
||||
return tool;
|
||||
}
|
||||
|
||||
describe("Tool Search", () => {
|
||||
it("enables object config when a mode is set", () => {
|
||||
expect(
|
||||
__testing.resolveToolSearchConfig({
|
||||
tools: {
|
||||
toolSearch: {
|
||||
mode: "tools",
|
||||
},
|
||||
},
|
||||
} as never),
|
||||
).toMatchObject({
|
||||
enabled: true,
|
||||
mode: "tools",
|
||||
});
|
||||
});
|
||||
|
||||
it("falls back to structured controls when code mode is unsupported", () => {
|
||||
__testing.setToolSearchCodeModeSupportedForTest(false);
|
||||
try {
|
||||
const config = { tools: { toolSearch: true } } as never;
|
||||
const resolved = __testing.resolveToolSearchConfig(config);
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: [
|
||||
fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode"),
|
||||
fakeTool(TOOL_SEARCH_RAW_TOOL_NAME, "search"),
|
||||
fakeTool(TOOL_DESCRIBE_RAW_TOOL_NAME, "describe"),
|
||||
fakeTool(TOOL_CALL_RAW_TOOL_NAME, "call"),
|
||||
pluginTool("fake_bun_fallback", "Fallback target"),
|
||||
],
|
||||
config,
|
||||
sessionId: "session-code-unsupported",
|
||||
});
|
||||
|
||||
expect(resolved.mode).toBe("tools");
|
||||
expect(compacted.tools.map((tool) => tool.name)).toEqual([
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
]);
|
||||
expect(compacted.catalogToolCount).toBe(1);
|
||||
} finally {
|
||||
__testing.setToolSearchCodeModeSupportedForTest(undefined);
|
||||
}
|
||||
});
|
||||
|
||||
it("compacts plugin tools behind the code surface and can search, describe, and call them", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const alpha = pluginTool("fake_create_ticket", "Create a ticket in the fake tracker");
|
||||
const beta = pluginTool("fake_weather", "Read fake weather");
|
||||
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: [codeTool, alpha, beta],
|
||||
config: {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
} as never,
|
||||
sessionId: "session-1",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
expect(compacted.tools.map((tool) => tool.name)).toEqual([TOOL_SEARCH_CODE_MODE_TOOL_NAME]);
|
||||
expect(compacted.catalogToolCount).toBe(2);
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-1",
|
||||
sessionKey: "agent:main:main",
|
||||
config: compacted.tools[0] ? {} : undefined,
|
||||
});
|
||||
const result = await runtimeCodeTool.execute("call-1", {
|
||||
code: `
|
||||
const hits = await openclaw.tools.search("ticket", { limit: 1 });
|
||||
const described = await openclaw.tools.describe(hits[0].id);
|
||||
return await openclaw.tools.call(described.id, { value: "ship" });
|
||||
`,
|
||||
});
|
||||
|
||||
expect(alpha.execute).toHaveBeenCalledWith(
|
||||
"tool_search_code:call-1:fake_create_ticket:1",
|
||||
{
|
||||
value: "ship",
|
||||
},
|
||||
expect.any(AbortSignal),
|
||||
undefined,
|
||||
undefined,
|
||||
);
|
||||
expect(result.details).toMatchObject({
|
||||
ok: true,
|
||||
telemetry: {
|
||||
catalogSize: 2,
|
||||
searchCount: 1,
|
||||
describeCount: 1,
|
||||
callCount: 1,
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("scopes catalogs by run id when attempts share a session", async () => {
|
||||
const runATool = pluginTool("fake_run_a", "Tool visible only to run A");
|
||||
const runBTool = pluginTool("fake_run_b", "Tool visible only to run B");
|
||||
const config = {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
} as never;
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode"), runATool],
|
||||
config,
|
||||
sessionId: "session-overlap",
|
||||
sessionKey: "agent:main:main",
|
||||
runId: "run-a",
|
||||
});
|
||||
applyToolSearchCatalog({
|
||||
tools: [fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode"), runBTool],
|
||||
config,
|
||||
sessionId: "session-overlap",
|
||||
sessionKey: "agent:main:main",
|
||||
runId: "run-b",
|
||||
});
|
||||
|
||||
const [, , , runACallTool] = createToolSearchTools({
|
||||
sessionId: "session-overlap",
|
||||
sessionKey: "agent:main:main",
|
||||
runId: "run-a",
|
||||
config,
|
||||
});
|
||||
await runACallTool.execute("call-run-a", {
|
||||
id: "fake_run_a",
|
||||
args: { value: "A" },
|
||||
});
|
||||
await expect(
|
||||
runACallTool.execute("call-run-a-miss", {
|
||||
id: "fake_run_b",
|
||||
args: { value: "B" },
|
||||
}),
|
||||
).rejects.toThrow("Unknown tool id: fake_run_b");
|
||||
|
||||
clearToolSearchCatalog({
|
||||
sessionId: "session-overlap",
|
||||
sessionKey: "agent:main:main",
|
||||
runId: "run-a",
|
||||
});
|
||||
expect(__testing.sessionCatalogs.has("run:run-a")).toBe(false);
|
||||
expect(__testing.sessionCatalogs.has("run:run-b")).toBe(true);
|
||||
expect(runATool.execute).toHaveBeenCalledTimes(1);
|
||||
expect(runBTool.execute).not.toHaveBeenCalled();
|
||||
clearToolSearchCatalog({ runId: "run-b" });
|
||||
});
|
||||
|
||||
it("uses the runtime-local catalog ref before the shared catalog registry", async () => {
|
||||
const localRef = createToolSearchCatalogRef();
|
||||
const localTool = pluginTool("fake_local_ref", "Tool visible through the local ref");
|
||||
const globalTool = pluginTool("fake_global_ref", "Tool visible through the registry fallback");
|
||||
const config = { tools: { toolSearch: true } } as never;
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode"), localTool],
|
||||
config,
|
||||
sessionId: "session-catalog-ref",
|
||||
runId: "run-local-ref",
|
||||
catalogRef: localRef,
|
||||
});
|
||||
applyToolSearchCatalog({
|
||||
tools: [fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode"), globalTool],
|
||||
config,
|
||||
sessionId: "session-catalog-ref",
|
||||
});
|
||||
|
||||
const [, , , callTool] = createToolSearchTools({
|
||||
sessionId: "session-catalog-ref",
|
||||
runId: "run-local-ref",
|
||||
catalogRef: localRef,
|
||||
config,
|
||||
});
|
||||
await callTool.execute("call-local-ref", {
|
||||
id: "fake_local_ref",
|
||||
args: { value: "local" },
|
||||
});
|
||||
await expect(
|
||||
callTool.execute("call-global-ref", {
|
||||
id: "fake_global_ref",
|
||||
args: { value: "global" },
|
||||
}),
|
||||
).rejects.toThrow("Unknown tool id: fake_global_ref");
|
||||
|
||||
expect(localTool.execute).toHaveBeenCalledTimes(1);
|
||||
expect(globalTool.execute).not.toHaveBeenCalled();
|
||||
clearToolSearchCatalog({ runId: "run-local-ref", catalogRef: localRef });
|
||||
clearToolSearchCatalog({ sessionId: "session-catalog-ref" });
|
||||
});
|
||||
|
||||
it("keeps raw fallback tools and hides the code tool in tools mode", () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const searchTool = fakeTool(TOOL_SEARCH_RAW_TOOL_NAME, "search");
|
||||
const describeTool = fakeTool(TOOL_DESCRIBE_RAW_TOOL_NAME, "describe");
|
||||
const callTool = fakeTool(TOOL_CALL_RAW_TOOL_NAME, "call");
|
||||
const target = pluginTool("fake_lookup", "Lookup fake records");
|
||||
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: [codeTool, searchTool, describeTool, callTool, target],
|
||||
config: {
|
||||
tools: {
|
||||
toolSearch: { enabled: true, mode: "tools" },
|
||||
},
|
||||
} as never,
|
||||
sessionId: "session-raw",
|
||||
});
|
||||
|
||||
expect(compacted.tools.map((tool) => tool.name)).toEqual([
|
||||
TOOL_SEARCH_RAW_TOOL_NAME,
|
||||
TOOL_DESCRIBE_RAW_TOOL_NAME,
|
||||
TOOL_CALL_RAW_TOOL_NAME,
|
||||
]);
|
||||
expect(compacted.catalogToolCount).toBe(1);
|
||||
});
|
||||
|
||||
it("drops inactive controls when the selected Tool Search control is unavailable", () => {
|
||||
const searchTool = fakeTool(TOOL_SEARCH_RAW_TOOL_NAME, "search");
|
||||
const describeTool = fakeTool(TOOL_DESCRIBE_RAW_TOOL_NAME, "describe");
|
||||
const callTool = fakeTool(TOOL_CALL_RAW_TOOL_NAME, "call");
|
||||
const target = pluginTool("fake_lookup_direct", "Lookup fake records directly");
|
||||
|
||||
const compacted = applyToolSearchCatalog({
|
||||
tools: [searchTool, describeTool, callTool, target],
|
||||
config: {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
} as never,
|
||||
sessionId: "session-code-control-denied",
|
||||
});
|
||||
|
||||
expect(compacted.tools.map((tool) => tool.name)).toEqual(["fake_lookup_direct"]);
|
||||
expect(compacted.catalogRegistered).toBe(false);
|
||||
expect(compacted.catalogToolCount).toBe(0);
|
||||
});
|
||||
|
||||
it("moves client tools into the same catalog when a session catalog exists", () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const config = {
|
||||
tools: {
|
||||
toolSearch: true,
|
||||
},
|
||||
} as never;
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool],
|
||||
config,
|
||||
sessionId: "session-client",
|
||||
});
|
||||
|
||||
const clientTool = fakeTool("client_pick_file", "Ask the client to pick a file");
|
||||
const compacted = addClientToolsToToolSearchCatalog({
|
||||
tools: [clientTool],
|
||||
config,
|
||||
sessionId: "session-client",
|
||||
});
|
||||
|
||||
expect(compacted.tools).toEqual([]);
|
||||
expect(compacted.catalogToolCount).toBe(1);
|
||||
expect(__testing.sessionCatalogs.get("session:session-client")?.entries).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
id: "client:client:client_pick_file",
|
||||
source: "client",
|
||||
}),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
it("wraps cataloged OpenClaw tools with before_tool_call hooks", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_hooked", "Run a hook-aware fake tool");
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-hooks",
|
||||
toolHookContext: {
|
||||
agentId: "agent-main",
|
||||
sessionId: "session-hooks",
|
||||
sessionKey: "agent:main:main",
|
||||
},
|
||||
});
|
||||
|
||||
const entry = __testing.sessionCatalogs
|
||||
.get("session:session-hooks")
|
||||
?.entries.find((candidate) => candidate.name === "fake_hooked");
|
||||
expect(entry).toBeTruthy();
|
||||
expect(isToolWrappedWithBeforeToolCallHook(entry!.tool as AnyAgentTool)).toBe(true);
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-hooks",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
await runtimeCodeTool.execute("call-hooks", {
|
||||
code: `return await openclaw.tools.call("fake_hooked", { value: "ok" });`,
|
||||
});
|
||||
expect(target.execute).toHaveBeenCalledWith(
|
||||
"tool_search_code:call-hooks:fake_hooked:1",
|
||||
{ value: "ok" },
|
||||
expect.any(AbortSignal),
|
||||
undefined,
|
||||
);
|
||||
});
|
||||
|
||||
it("does not re-wrap abort-wrapped tools that already have before_tool_call hooks", () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_already_hooked", "Already hook-aware fake tool");
|
||||
const hooked = wrapToolWithBeforeToolCallHook(target, {
|
||||
agentId: "agent-main",
|
||||
sessionId: "session-hooks-abort",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
const abortWrapped = wrapToolWithAbortSignal(hooked, new AbortController().signal);
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, abortWrapped],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-hooks-abort",
|
||||
toolHookContext: {
|
||||
agentId: "agent-main",
|
||||
sessionId: "session-hooks-abort",
|
||||
sessionKey: "agent:main:main",
|
||||
},
|
||||
});
|
||||
|
||||
const entry = __testing.sessionCatalogs
|
||||
.get("session:session-hooks-abort")
|
||||
?.entries.find((candidate) => candidate.name === "fake_already_hooked");
|
||||
expect(entry?.tool).toBe(abortWrapped);
|
||||
expect(isToolWrappedWithBeforeToolCallHook(entry!.tool as AnyAgentTool)).toBe(true);
|
||||
});
|
||||
|
||||
it("uses a unique bridged tool call id for repeated calls", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_repeated", "Run a repeated fake tool");
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-repeated",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-repeated",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
await runtimeCodeTool.execute("call-repeated", {
|
||||
code: `
|
||||
await openclaw.tools.call("fake_repeated", { value: "one" });
|
||||
return await openclaw.tools.call("fake_repeated", { value: "two" });
|
||||
`,
|
||||
});
|
||||
|
||||
expect(target.execute).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
"tool_search_code:call-repeated:fake_repeated:1",
|
||||
{
|
||||
value: "one",
|
||||
},
|
||||
expect.any(AbortSignal),
|
||||
undefined,
|
||||
undefined,
|
||||
);
|
||||
expect(target.execute).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
"tool_search_code:call-repeated:fake_repeated:2",
|
||||
{
|
||||
value: "two",
|
||||
},
|
||||
expect.any(AbortSignal),
|
||||
undefined,
|
||||
undefined,
|
||||
);
|
||||
await runtimeCodeTool.execute("call-repeated-again", {
|
||||
code: `return await openclaw.tools.call("fake_repeated", { value: "three" });`,
|
||||
});
|
||||
|
||||
expect(target.execute).toHaveBeenNthCalledWith(
|
||||
3,
|
||||
"tool_search_code:call-repeated-again:fake_repeated:1",
|
||||
{
|
||||
value: "three",
|
||||
},
|
||||
expect.any(AbortSignal),
|
||||
undefined,
|
||||
undefined,
|
||||
);
|
||||
});
|
||||
|
||||
it("routes bridged calls through the configured catalog executor", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_lifecycle", "Run through lifecycle executor");
|
||||
const abortController = new AbortController();
|
||||
const onUpdate = vi.fn();
|
||||
const executeTool = vi.fn(async () => jsonResult({ status: "ok" }));
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-lifecycle",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool, , , runtimeCallTool] = createToolSearchTools({
|
||||
sessionId: "session-lifecycle",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
abortSignal: abortController.signal,
|
||||
executeTool,
|
||||
});
|
||||
await runtimeCodeTool.execute(
|
||||
"call-lifecycle",
|
||||
{
|
||||
code: `return await openclaw.tools.call("fake_lifecycle", { value: "ok" });`,
|
||||
},
|
||||
undefined,
|
||||
onUpdate,
|
||||
);
|
||||
|
||||
expect(target.execute).not.toHaveBeenCalled();
|
||||
expect(executeTool).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
tool: expect.objectContaining({ name: "fake_lifecycle" }),
|
||||
toolName: "fake_lifecycle",
|
||||
toolCallId: "tool_search_code:call-lifecycle:fake_lifecycle:1",
|
||||
parentToolCallId: "call-lifecycle",
|
||||
input: { value: "ok" },
|
||||
signal: expect.any(AbortSignal),
|
||||
onUpdate,
|
||||
}),
|
||||
);
|
||||
|
||||
await runtimeCallTool.execute(
|
||||
"call-lifecycle-structured",
|
||||
{
|
||||
id: "fake_lifecycle",
|
||||
args: { value: "structured" },
|
||||
},
|
||||
abortController.signal,
|
||||
onUpdate,
|
||||
);
|
||||
|
||||
expect(target.execute).not.toHaveBeenCalled();
|
||||
expect(executeTool).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
expect.objectContaining({
|
||||
tool: expect.objectContaining({ name: "fake_lifecycle" }),
|
||||
toolName: "fake_lifecycle",
|
||||
toolCallId: "tool_search_code:call-lifecycle-structured:fake_lifecycle:1",
|
||||
parentToolCallId: "call-lifecycle-structured",
|
||||
input: { value: "structured" },
|
||||
signal: abortController.signal,
|
||||
onUpdate,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("projects target tool calls after their Tool Search wrapper result", () => {
|
||||
const messages = [
|
||||
{
|
||||
role: "assistant",
|
||||
content: [
|
||||
{
|
||||
type: "toolCall",
|
||||
id: "wrapper-call",
|
||||
name: TOOL_CALL_RAW_TOOL_NAME,
|
||||
arguments: { id: "fake_target", args: { value: "ok" } },
|
||||
},
|
||||
],
|
||||
},
|
||||
{
|
||||
role: "toolResult",
|
||||
toolCallId: "wrapper-call",
|
||||
toolName: TOOL_CALL_RAW_TOOL_NAME,
|
||||
content: [{ type: "text", text: "wrapped" }],
|
||||
},
|
||||
{
|
||||
role: "assistant",
|
||||
content: [{ type: "text", text: "done" }],
|
||||
},
|
||||
];
|
||||
|
||||
const projected = projectToolSearchTargetTranscriptMessages(messages as never, [
|
||||
{
|
||||
parentToolCallId: "wrapper-call",
|
||||
toolCallId: "tool_search_code:wrapper-call:fake_target:1",
|
||||
toolName: "fake_target",
|
||||
input: { value: "ok" },
|
||||
result: jsonResult({ ok: true }),
|
||||
timestamp: 123,
|
||||
},
|
||||
]);
|
||||
|
||||
expect(projected).toHaveLength(5);
|
||||
expect(projected[2]).toMatchObject({
|
||||
role: "assistant",
|
||||
content: [
|
||||
{
|
||||
type: "toolCall",
|
||||
id: "tool_search_code:wrapper-call:fake_target:1",
|
||||
name: "fake_target",
|
||||
arguments: { value: "ok" },
|
||||
input: { value: "ok" },
|
||||
},
|
||||
],
|
||||
});
|
||||
expect(projected[3]).toMatchObject({
|
||||
role: "toolResult",
|
||||
toolCallId: "tool_search_code:wrapper-call:fake_target:1",
|
||||
toolName: "fake_target",
|
||||
isError: false,
|
||||
content: [{ type: "text", text: JSON.stringify({ ok: true }, null, 2) }],
|
||||
});
|
||||
expect(projected[4]).toBe(messages[2]);
|
||||
});
|
||||
|
||||
it("does not execute fire-and-forget bridged calls after code returns", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_fire_and_forget", "Should not run unless awaited");
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-fire-and-forget",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-fire-and-forget",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
const result = await runtimeCodeTool.execute("call-fire-and-forget", {
|
||||
code: `
|
||||
openclaw.tools.call("fake_fire_and_forget", { value: "late" });
|
||||
return "done";
|
||||
`,
|
||||
});
|
||||
|
||||
expect(target.execute).not.toHaveBeenCalled();
|
||||
expect(result.details).toMatchObject({
|
||||
ok: true,
|
||||
value: "done",
|
||||
telemetry: {
|
||||
callCount: 0,
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("waits for started bridged calls before returning code-mode success", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_then_started", "Started by .then without await");
|
||||
let resolveTool: (() => void) | undefined;
|
||||
target.execute = vi.fn(
|
||||
async (_toolCallId: string, input: unknown): Promise<ReturnType<typeof jsonResult>> => {
|
||||
await new Promise<void>((resolve) => {
|
||||
resolveTool = resolve;
|
||||
});
|
||||
return jsonResult({ name: target.name, input });
|
||||
},
|
||||
);
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-started-bridge",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-started-bridge",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
let settled = false;
|
||||
const resultPromise = runtimeCodeTool
|
||||
.execute("call-started-bridge", {
|
||||
code: `
|
||||
openclaw.tools.call("fake_then_started", { value: "started" }).then(() => {});
|
||||
return "done";
|
||||
`,
|
||||
})
|
||||
.then((result) => {
|
||||
settled = true;
|
||||
return result;
|
||||
});
|
||||
|
||||
await vi.waitFor(() => expect(target.execute).toHaveBeenCalledTimes(1));
|
||||
await new Promise((resolve) => setTimeout(resolve, 20));
|
||||
expect(settled).toBe(false);
|
||||
resolveTool?.();
|
||||
const result = await resultPromise;
|
||||
|
||||
expect(result.details).toMatchObject({
|
||||
ok: true,
|
||||
value: "done",
|
||||
telemetry: {
|
||||
callCount: 1,
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("does not expose the host process to model-authored code", async () => {
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-escape",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-escape", {
|
||||
code: `return Function("return process")();`,
|
||||
}),
|
||||
).rejects.toThrow();
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-constructor-escape", {
|
||||
code: `return globalThis.constructor.constructor("return process")();`,
|
||||
}),
|
||||
).rejects.toThrow();
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-console-escape", {
|
||||
code: `return console.log.constructor.constructor("return process")();`,
|
||||
}),
|
||||
).rejects.toThrow();
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-bridge-escape", {
|
||||
code: `return openclaw.tools.call.constructor.constructor("return process")();`,
|
||||
}),
|
||||
).rejects.toThrow();
|
||||
});
|
||||
|
||||
it("preserves code-mode bridge errors from the child process", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-missing-tool-error",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-missing-tool-error",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-missing-tool", {
|
||||
code: `return await openclaw.tools.call("missing_tool", {});`,
|
||||
}),
|
||||
).rejects.toThrow("Unknown tool id: missing_tool");
|
||||
});
|
||||
|
||||
it("does not expose host-realm bridge result objects to model-authored code", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_bridge_result_escape", "Target for bridge result escape");
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-bridge-result-escape",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-bridge-result-escape",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-bridge-result-escape", {
|
||||
code: `
|
||||
const hits = await openclaw.tools.search("bridge result", { limit: 1 });
|
||||
return hits.constructor.constructor("return process")();
|
||||
`,
|
||||
}),
|
||||
).rejects.toThrow();
|
||||
expect(target.execute).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it("does not let model-authored code access bridge controller locals", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_controller_escape", "Target for forged bridge request");
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config: { tools: { toolSearch: true } } as never,
|
||||
sessionId: "session-controller-escape",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-controller-escape",
|
||||
sessionKey: "agent:main:main",
|
||||
config: {},
|
||||
});
|
||||
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-controller-escape", {
|
||||
code: `
|
||||
})(openclaw, console),
|
||||
bridgeMessages.push({
|
||||
id: "forged",
|
||||
method: "call",
|
||||
args: ["fake_controller_escape", { value: "forged" }],
|
||||
}),
|
||||
(async (openclaw, console) => {
|
||||
return "done";
|
||||
`,
|
||||
}),
|
||||
).rejects.toThrow();
|
||||
expect(target.execute).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it("terminates async continuations that block the event loop after a bridge call", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const alpha = pluginTool("fake_timeout_target", "Target tool for timeout search");
|
||||
|
||||
const config = {
|
||||
tools: {
|
||||
toolSearch: { enabled: true, mode: "code", codeTimeoutMs: 1000 },
|
||||
},
|
||||
} as never;
|
||||
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, alpha],
|
||||
config,
|
||||
sessionId: "session-timeout",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-timeout",
|
||||
sessionKey: "agent:main:main",
|
||||
config,
|
||||
});
|
||||
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-timeout", {
|
||||
code: `
|
||||
await openclaw.tools.search("timeout", { limit: 1 });
|
||||
while (true) {}
|
||||
`,
|
||||
}),
|
||||
).rejects.toThrow("tool_search_code timed out");
|
||||
}, 5_000);
|
||||
|
||||
it("aborts already-started bridged calls when code mode times out", async () => {
|
||||
const codeTool = fakeTool(TOOL_SEARCH_CODE_MODE_TOOL_NAME, "code mode");
|
||||
const target = pluginTool("fake_abort_on_timeout", "Long-running target tool");
|
||||
let observedSignal: AbortSignal | undefined;
|
||||
let abortCount = 0;
|
||||
target.execute = vi.fn(
|
||||
async (
|
||||
_toolCallId: string,
|
||||
_input: unknown,
|
||||
signal?: AbortSignal,
|
||||
): Promise<ReturnType<typeof jsonResult>> => {
|
||||
observedSignal = signal;
|
||||
await new Promise<void>((resolve) => {
|
||||
if (signal?.aborted) {
|
||||
abortCount += 1;
|
||||
resolve();
|
||||
return;
|
||||
}
|
||||
signal?.addEventListener(
|
||||
"abort",
|
||||
() => {
|
||||
abortCount += 1;
|
||||
resolve();
|
||||
},
|
||||
{ once: true },
|
||||
);
|
||||
});
|
||||
return jsonResult({ aborted: true });
|
||||
},
|
||||
);
|
||||
|
||||
const config = {
|
||||
tools: {
|
||||
toolSearch: { enabled: true, mode: "code", codeTimeoutMs: 100 },
|
||||
},
|
||||
} as never;
|
||||
applyToolSearchCatalog({
|
||||
tools: [codeTool, target],
|
||||
config,
|
||||
sessionId: "session-abort-timeout",
|
||||
sessionKey: "agent:main:main",
|
||||
});
|
||||
|
||||
const [runtimeCodeTool] = createToolSearchTools({
|
||||
sessionId: "session-abort-timeout",
|
||||
sessionKey: "agent:main:main",
|
||||
config,
|
||||
});
|
||||
|
||||
await expect(
|
||||
runtimeCodeTool.execute("call-abort-timeout", {
|
||||
code: `return await openclaw.tools.call("fake_abort_on_timeout", { value: "wait" });`,
|
||||
}),
|
||||
).rejects.toThrow("tool_search_code timed out");
|
||||
expect(observedSignal).toBeDefined();
|
||||
expect(observedSignal?.aborted).toBe(true);
|
||||
expect(abortCount).toBe(1);
|
||||
});
|
||||
});
|
||||
File diff suppressed because it is too large
Load Diff
@@ -120,43 +120,6 @@ function requireImageGenerateTool(tool: ReturnType<typeof createImageGenerateToo
|
||||
return tool;
|
||||
}
|
||||
|
||||
type UnknownMock = { mock: { calls: unknown[][] } };
|
||||
|
||||
function mockCallArg(
|
||||
mock: unknown,
|
||||
index: number,
|
||||
label: string,
|
||||
argIndex = 0,
|
||||
): Record<string, unknown> {
|
||||
const calls = (mock as UnknownMock).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`Expected ${label} to be a mock`);
|
||||
}
|
||||
const call = calls[index];
|
||||
if (!call) {
|
||||
throw new Error(`Expected ${label} call ${index + 1}`);
|
||||
}
|
||||
return call[argIndex] as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object") {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
type ImageGenerateTool = NonNullable<ReturnType<typeof createImageGenerateTool>>;
|
||||
type ToolResult = Awaited<ReturnType<ImageGenerateTool["execute"]>>;
|
||||
|
||||
function resultDetails(result: ToolResult): Record<string, unknown> {
|
||||
return requireRecord(result.details, "tool result details");
|
||||
}
|
||||
|
||||
function resultText(result: ToolResult): string {
|
||||
return (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
}
|
||||
|
||||
function ensureDefaultImageGenerationProvidersStubbed() {
|
||||
if (vi.isMockFunction(imageGenerationRuntime.listRuntimeImageGenerationProviders)) {
|
||||
return;
|
||||
@@ -609,23 +572,26 @@ describe("createImageGenerateTool", () => {
|
||||
size: "1024x1024",
|
||||
});
|
||||
|
||||
const generateArgs = mockCallArg(generateImage, 0, "generateImage");
|
||||
expect(generateArgs.cfg).toEqual({
|
||||
agents: {
|
||||
defaults: {
|
||||
mediaMaxMb: 8,
|
||||
imageGenerationModel: {
|
||||
primary: "openai/gpt-image-1",
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
cfg: {
|
||||
agents: {
|
||||
defaults: {
|
||||
mediaMaxMb: 8,
|
||||
imageGenerationModel: {
|
||||
primary: "openai/gpt-image-1",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
});
|
||||
expect(generateArgs.prompt).toBe("A cat wearing sunglasses");
|
||||
expect(generateArgs.agentDir).toBe("/tmp/agent");
|
||||
expect(generateArgs.modelOverride).toBe("openai/gpt-image-1");
|
||||
expect(generateArgs.size).toBe("1024x1024");
|
||||
expect(generateArgs.count).toBe(2);
|
||||
expect(generateArgs.inputImages).toEqual([]);
|
||||
prompt: "A cat wearing sunglasses",
|
||||
agentDir: "/tmp/agent",
|
||||
modelOverride: "openai/gpt-image-1",
|
||||
size: "1024x1024",
|
||||
count: 2,
|
||||
inputImages: [],
|
||||
}),
|
||||
);
|
||||
expect(saveMediaBuffer).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
Buffer.from("png-1"),
|
||||
@@ -642,17 +608,26 @@ describe("createImageGenerateTool", () => {
|
||||
8 * 1024 * 1024,
|
||||
"cats/output.png",
|
||||
);
|
||||
const text = resultText(result);
|
||||
expect(text).toContain("Generated 2 images with openai/gpt-image-1.");
|
||||
const details = resultDetails(result);
|
||||
const media = requireRecord(details.media, "media details");
|
||||
expect(details.provider).toBe("openai");
|
||||
expect(details.model).toBe("gpt-image-1");
|
||||
expect(details.count).toBe(2);
|
||||
expect(media.mediaUrls).toEqual(["/tmp/generated-1.png", "/tmp/generated-2.png"]);
|
||||
expect(details.paths).toEqual(["/tmp/generated-1.png", "/tmp/generated-2.png"]);
|
||||
expect(details.filename).toBe("cats/output.png");
|
||||
expect(details.revisedPrompts).toEqual(["A more cinematic cat"]);
|
||||
expect(result).toMatchObject({
|
||||
content: [
|
||||
{
|
||||
type: "text",
|
||||
text: expect.stringContaining("Generated 2 images with openai/gpt-image-1."),
|
||||
},
|
||||
],
|
||||
details: {
|
||||
provider: "openai",
|
||||
model: "gpt-image-1",
|
||||
count: 2,
|
||||
media: {
|
||||
mediaUrls: ["/tmp/generated-1.png", "/tmp/generated-2.png"],
|
||||
},
|
||||
paths: ["/tmp/generated-1.png", "/tmp/generated-2.png"],
|
||||
filename: "cats/output.png",
|
||||
revisedPrompts: ["A more cinematic cat"],
|
||||
},
|
||||
});
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
expect(text).toContain("MEDIA:/tmp/generated-1.png");
|
||||
expect(text).toContain("MEDIA:/tmp/generated-2.png");
|
||||
});
|
||||
@@ -702,10 +677,20 @@ describe("createImageGenerateTool", () => {
|
||||
timeoutMs: 12_345,
|
||||
});
|
||||
|
||||
expect(mockCallArg(generateImage, 0, "generateImage").timeoutMs).toBe(180_000);
|
||||
expect(mockCallArg(generateImage, 1, "generateImage").timeoutMs).toBe(12_345);
|
||||
expect(resultDetails(defaultResult).timeoutMs).toBe(180_000);
|
||||
expect(resultDetails(overrideResult).timeoutMs).toBe(12_345);
|
||||
expect(generateImage).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
expect.objectContaining({
|
||||
timeoutMs: 180_000,
|
||||
}),
|
||||
);
|
||||
expect(generateImage).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
expect.objectContaining({
|
||||
timeoutMs: 12_345,
|
||||
}),
|
||||
);
|
||||
expect(defaultResult.details).toMatchObject({ timeoutMs: 180_000 });
|
||||
expect(overrideResult.details).toMatchObject({ timeoutMs: 12_345 });
|
||||
});
|
||||
|
||||
it("forwards output hints and OpenAI provider options", async () => {
|
||||
@@ -742,20 +727,26 @@ describe("createImageGenerateTool", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const generateArgs = mockCallArg(generateImage, 0, "generateImage");
|
||||
expect(generateArgs.quality).toBe("low");
|
||||
expect(generateArgs.outputFormat).toBe("jpeg");
|
||||
expect(generateArgs.providerOptions).toEqual({
|
||||
openai: {
|
||||
background: "opaque",
|
||||
moderation: "low",
|
||||
outputCompression: 60,
|
||||
user: "end-user-42",
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
quality: "low",
|
||||
outputFormat: "jpeg",
|
||||
providerOptions: {
|
||||
openai: {
|
||||
background: "opaque",
|
||||
moderation: "low",
|
||||
outputCompression: 60,
|
||||
user: "end-user-42",
|
||||
},
|
||||
},
|
||||
}),
|
||||
);
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
quality: "low",
|
||||
outputFormat: "jpeg",
|
||||
},
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.quality).toBe("low");
|
||||
expect(details.outputFormat).toBe("jpeg");
|
||||
});
|
||||
|
||||
it("forwards transparent OpenAI background requests with a PNG output format", async () => {
|
||||
@@ -788,21 +779,30 @@ describe("createImageGenerateTool", () => {
|
||||
},
|
||||
});
|
||||
|
||||
const generateArgs = mockCallArg(generateImage, 0, "generateImage");
|
||||
const cfg = requireRecord(generateArgs.cfg, "generateImage config");
|
||||
const agents = requireRecord(cfg.agents, "generateImage agents config");
|
||||
const defaults = requireRecord(agents.defaults, "generateImage defaults config");
|
||||
expect(defaults.imageGenerationModel).toEqual({ primary: "openai/gpt-image-1.5" });
|
||||
expect(generateArgs.outputFormat).toBe("png");
|
||||
expect(generateArgs.providerOptions).toEqual({
|
||||
openai: {
|
||||
background: "transparent",
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
cfg: expect.objectContaining({
|
||||
agents: expect.objectContaining({
|
||||
defaults: expect.objectContaining({
|
||||
imageGenerationModel: { primary: "openai/gpt-image-1.5" },
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
outputFormat: "png",
|
||||
providerOptions: {
|
||||
openai: {
|
||||
background: "transparent",
|
||||
},
|
||||
},
|
||||
}),
|
||||
);
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
provider: "openai",
|
||||
model: "gpt-image-1.5",
|
||||
outputFormat: "png",
|
||||
},
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.provider).toBe("openai");
|
||||
expect(details.model).toBe("gpt-image-1.5");
|
||||
expect(details.outputFormat).toBe("png");
|
||||
});
|
||||
|
||||
it("includes MEDIA paths in content text so follow-up replies use the real saved file", async () => {
|
||||
@@ -866,16 +866,18 @@ describe("createImageGenerateTool", () => {
|
||||
);
|
||||
|
||||
const result = await tool.execute("call-regression", { prompt: "kodo sawaki zazen" });
|
||||
const text = resultText(result);
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
|
||||
expect(text).toContain(
|
||||
"MEDIA:/home/openclaw/.openclaw/media/tool-image-generation/kodo_sawaki_zazen---3337a0ed-898a-4572-8950-0d288719f4f8.jpg",
|
||||
);
|
||||
const details = resultDetails(result);
|
||||
const media = requireRecord(details.media, "media details");
|
||||
expect(media.mediaUrls).toEqual([
|
||||
"/home/openclaw/.openclaw/media/tool-image-generation/kodo_sawaki_zazen---3337a0ed-898a-4572-8950-0d288719f4f8.jpg",
|
||||
]);
|
||||
expect(result.details).toMatchObject({
|
||||
media: {
|
||||
mediaUrls: [
|
||||
"/home/openclaw/.openclaw/media/tool-image-generation/kodo_sawaki_zazen---3337a0ed-898a-4572-8950-0d288719f4f8.jpg",
|
||||
],
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("rejects counts outside the supported range", async () => {
|
||||
@@ -936,15 +938,18 @@ describe("createImageGenerateTool", () => {
|
||||
image: "./fixtures/reference.png",
|
||||
});
|
||||
|
||||
const generateArgs = mockCallArg(generateImage, 0, "generateImage");
|
||||
expect(generateArgs.aspectRatio).toBeUndefined();
|
||||
expect(generateArgs.resolution).toBe("4K");
|
||||
expect(generateArgs.inputImages).toEqual([
|
||||
{
|
||||
buffer: Buffer.from("input-image"),
|
||||
mimeType: "image/png",
|
||||
},
|
||||
]);
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
aspectRatio: undefined,
|
||||
resolution: "4K",
|
||||
inputImages: [
|
||||
expect.objectContaining({
|
||||
buffer: Buffer.from("input-image"),
|
||||
mimeType: "image/png",
|
||||
}),
|
||||
],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("accepts managed inbound reference images for edit mode", async () => {
|
||||
@@ -958,12 +963,10 @@ describe("createImageGenerateTool", () => {
|
||||
image: "media://inbound/reference.png",
|
||||
});
|
||||
|
||||
const loadArgs = mockCallArg(webMedia.loadWebMedia, 0, "loadWebMedia", 1);
|
||||
expect(mockCallArg(webMedia.loadWebMedia, 0, "loadWebMedia", 0)).toBe(
|
||||
expect(webMedia.loadWebMedia).toHaveBeenCalledWith(
|
||||
"media://inbound/reference.png",
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(loadArgs).not.toBeNull();
|
||||
expect(typeof loadArgs).toBe("object");
|
||||
});
|
||||
|
||||
it("passes web_fetch SSRF policy to remote reference images", async () => {
|
||||
@@ -984,10 +987,10 @@ describe("createImageGenerateTool", () => {
|
||||
prompt: "Use this reference.",
|
||||
image: "http://198.18.0.153/reference.png",
|
||||
});
|
||||
const defaultLoadUrl = mockCallArg(webMedia.loadWebMedia, 0, "loadWebMedia", 0);
|
||||
const defaultLoadOptions = mockCallArg(webMedia.loadWebMedia, 0, "loadWebMedia", 1);
|
||||
expect(defaultLoadUrl).toBe("http://198.18.0.153/reference.png");
|
||||
expect(requireRecord(defaultLoadOptions, "loadWebMedia options").ssrfPolicy).toBeUndefined();
|
||||
expect(webMedia.loadWebMedia).toHaveBeenLastCalledWith(
|
||||
"http://198.18.0.153/reference.png",
|
||||
expect.not.objectContaining({ ssrfPolicy: expect.anything() }),
|
||||
);
|
||||
|
||||
const tool = requireImageGenerateTool(
|
||||
createImageGenerateTool({
|
||||
@@ -1006,15 +1009,17 @@ describe("createImageGenerateTool", () => {
|
||||
image: "http://198.18.0.153/reference.png",
|
||||
});
|
||||
|
||||
const configuredLoadUrl = mockCallArg(webMedia.loadWebMedia, 1, "loadWebMedia", 0);
|
||||
const configuredLoadOptions = mockCallArg(webMedia.loadWebMedia, 1, "loadWebMedia", 1);
|
||||
expect(configuredLoadUrl).toBe("http://198.18.0.153/reference.png");
|
||||
expect(requireRecord(configuredLoadOptions, "loadWebMedia options").ssrfPolicy).toEqual({
|
||||
allowRfc2544BenchmarkRange: true,
|
||||
});
|
||||
expect(mockCallArg(generateImage, 1, "generateImage").ssrfPolicy).toEqual({
|
||||
allowRfc2544BenchmarkRange: true,
|
||||
});
|
||||
expect(webMedia.loadWebMedia).toHaveBeenCalledWith(
|
||||
"http://198.18.0.153/reference.png",
|
||||
expect.objectContaining({
|
||||
ssrfPolicy: { allowRfc2544BenchmarkRange: true },
|
||||
}),
|
||||
);
|
||||
expect(generateImage).toHaveBeenLastCalledWith(
|
||||
expect.objectContaining({
|
||||
ssrfPolicy: { allowRfc2544BenchmarkRange: true },
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("ignores non-finite mediaMaxMb when loading reference images", async () => {
|
||||
@@ -1041,10 +1046,9 @@ describe("createImageGenerateTool", () => {
|
||||
image: "./fixtures/reference.png",
|
||||
});
|
||||
|
||||
expect(typeof mockCallArg(webMedia.loadWebMedia, 0, "loadWebMedia", 0)).toBe("string");
|
||||
expect(mockCallArg(webMedia.loadWebMedia, 0, "loadWebMedia", 1)).toHaveProperty(
|
||||
"maxBytes",
|
||||
undefined,
|
||||
expect(webMedia.loadWebMedia).toHaveBeenCalledWith(
|
||||
expect.any(String),
|
||||
expect.objectContaining({ maxBytes: undefined }),
|
||||
);
|
||||
});
|
||||
|
||||
@@ -1115,19 +1119,25 @@ describe("createImageGenerateTool", () => {
|
||||
prompt: "Remove the subject but keep the rest unchanged.",
|
||||
image: "./fixtures/reference.png",
|
||||
});
|
||||
const details = resultDetails(result);
|
||||
expect(details.provider).toBe("openai");
|
||||
expect(details.model).toBe("gpt-image-1");
|
||||
|
||||
const generateArgs = mockCallArg(generateImage, 0, "generateImage");
|
||||
expect(generateArgs.modelOverride).toBeUndefined();
|
||||
expect(generateArgs.resolution).toBeUndefined();
|
||||
expect(generateArgs.inputImages).toEqual([
|
||||
{
|
||||
buffer: Buffer.from("input-image"),
|
||||
mimeType: "image/jpeg",
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
provider: "openai",
|
||||
model: "gpt-image-1",
|
||||
},
|
||||
]);
|
||||
});
|
||||
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
modelOverride: undefined,
|
||||
resolution: undefined,
|
||||
inputImages: [
|
||||
expect.objectContaining({
|
||||
buffer: Buffer.from("input-image"),
|
||||
mimeType: "image/jpeg",
|
||||
}),
|
||||
],
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("forwards explicit aspect ratio and supports up to 5 reference images", async () => {
|
||||
@@ -1143,15 +1153,16 @@ describe("createImageGenerateTool", () => {
|
||||
aspectRatio: "16:9",
|
||||
});
|
||||
|
||||
const generateArgs = mockCallArg(generateImage, 0, "generateImage");
|
||||
expect(generateArgs.autoProviderFallback).toBe(false);
|
||||
expect(generateArgs.aspectRatio).toBe("16:9");
|
||||
const inputImages = generateArgs.inputImages as Array<{ buffer: Buffer; mimeType: string }>;
|
||||
expect(inputImages).toHaveLength(5);
|
||||
expect(inputImages[0]).toEqual({
|
||||
buffer: Buffer.from("input-image"),
|
||||
mimeType: "image/png",
|
||||
});
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
autoProviderFallback: false,
|
||||
aspectRatio: "16:9",
|
||||
inputImages: expect.arrayContaining([
|
||||
expect.objectContaining({ buffer: Buffer.from("input-image"), mimeType: "image/png" }),
|
||||
]),
|
||||
}),
|
||||
);
|
||||
expect(generateImage.mock.calls[0]?.[0].inputImages).toHaveLength(5);
|
||||
});
|
||||
|
||||
it("reports ignored unsupported overrides instead of failing", async () => {
|
||||
@@ -1209,17 +1220,18 @@ describe("createImageGenerateTool", () => {
|
||||
prompt: "A lobster at the movies",
|
||||
aspectRatio: "1:1",
|
||||
});
|
||||
const text = resultText(result);
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
|
||||
expect(text).toContain("Generated 1 image with openai/gpt-image-1.");
|
||||
expect(text).toContain(
|
||||
"Warning: Ignored unsupported overrides for openai/gpt-image-1: aspectRatio=1:1.",
|
||||
);
|
||||
const details = resultDetails(result);
|
||||
expect(details.warning).toBe(
|
||||
"Ignored unsupported overrides for openai/gpt-image-1: aspectRatio=1:1.",
|
||||
);
|
||||
expect(details.ignoredOverrides).toEqual([{ key: "aspectRatio", value: "1:1" }]);
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
warning: "Ignored unsupported overrides for openai/gpt-image-1: aspectRatio=1:1.",
|
||||
ignoredOverrides: [{ key: "aspectRatio", value: "1:1" }],
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("surfaces normalized image geometry from runtime metadata", async () => {
|
||||
@@ -1259,19 +1271,20 @@ describe("createImageGenerateTool", () => {
|
||||
size: "1280x720",
|
||||
});
|
||||
|
||||
const details = resultDetails(result);
|
||||
expect(details.aspectRatio).toBe("16:9");
|
||||
expect(details.normalization).toEqual({
|
||||
aspectRatio: {
|
||||
applied: "16:9",
|
||||
derivedFrom: "size",
|
||||
expect(result.details).toMatchObject({
|
||||
aspectRatio: "16:9",
|
||||
normalization: {
|
||||
aspectRatio: {
|
||||
applied: "16:9",
|
||||
derivedFrom: "size",
|
||||
},
|
||||
},
|
||||
metadata: {
|
||||
requestedSize: "1280x720",
|
||||
normalizedAspectRatio: "16:9",
|
||||
},
|
||||
});
|
||||
expect(details.metadata).toEqual({
|
||||
requestedSize: "1280x720",
|
||||
normalizedAspectRatio: "16:9",
|
||||
});
|
||||
expect(details).not.toHaveProperty("size");
|
||||
expect(result.details).not.toHaveProperty("size");
|
||||
});
|
||||
|
||||
it("escapes image-generation summary text before appending tool MEDIA output", async () => {
|
||||
@@ -1328,7 +1341,7 @@ describe("createImageGenerateTool", () => {
|
||||
const result = await tool.execute("call-openai-generate", {
|
||||
prompt: "A lobster at the movies",
|
||||
});
|
||||
const text = resultText(result);
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
const parsed = splitMediaFromOutput(text);
|
||||
|
||||
expect(text).toContain(
|
||||
@@ -1336,12 +1349,13 @@ describe("createImageGenerateTool", () => {
|
||||
);
|
||||
expect(text).toContain("size=1024x1024\\nMEDIA:/etc/passwd\\t\\u2028\\u0000");
|
||||
expect(parsed.mediaUrls).toEqual(["/tmp/generated.png"]);
|
||||
const details = resultDetails(result);
|
||||
expect(details.provider).toBe("openai\nMEDIA:/tmp/provider.png");
|
||||
expect(details.model).toBe("gpt-image-1\nMEDIA:/etc/model.png");
|
||||
expect(details.ignoredOverrides).toEqual([
|
||||
{ key: "size", value: "1024x1024\nMEDIA:/etc/passwd\t\u2028\0" },
|
||||
]);
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
provider: "openai\nMEDIA:/tmp/provider.png",
|
||||
model: "gpt-image-1\nMEDIA:/etc/model.png",
|
||||
ignoredOverrides: [{ key: "size", value: "1024x1024\nMEDIA:/etc/passwd\t\u2028\0" }],
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("rejects unsupported aspect ratios", async () => {
|
||||
@@ -1386,7 +1400,7 @@ describe("createImageGenerateTool", () => {
|
||||
);
|
||||
|
||||
const result = await tool.execute("call-list", { action: "list" });
|
||||
const text = resultText(result);
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
|
||||
expect(text).toContain("google (default gemini-3.1-flash-image-preview)");
|
||||
expect(text).toContain("gemini-3.1-flash-image-preview");
|
||||
@@ -1397,27 +1411,31 @@ describe("createImageGenerateTool", () => {
|
||||
);
|
||||
expect(text).toContain("editing up to 5 refs");
|
||||
expect(text).toContain("aspect ratios 1:1, 16:9");
|
||||
const details = resultDetails(result);
|
||||
const providers = details.providers as Array<Record<string, unknown>>;
|
||||
const googleProvider = providers.find((provider) => provider.id === "google");
|
||||
const openaiProvider = providers.find((provider) => provider.id === "openai");
|
||||
if (!googleProvider || !openaiProvider) {
|
||||
throw new Error("Expected google and openai provider details");
|
||||
}
|
||||
expect(googleProvider.defaultModel).toBe("gemini-3.1-flash-image-preview");
|
||||
expect(googleProvider.authEnvVars).toEqual(["GEMINI_API_KEY", "GOOGLE_API_KEY"]);
|
||||
expect(googleProvider.models).toEqual([
|
||||
"gemini-3.1-flash-image-preview",
|
||||
"gemini-3-pro-image-preview",
|
||||
]);
|
||||
const googleCapabilities = requireRecord(googleProvider.capabilities, "google capabilities");
|
||||
expect(googleCapabilities.edit).toEqual({
|
||||
enabled: true,
|
||||
maxInputImages: 5,
|
||||
supportsAspectRatio: true,
|
||||
supportsResolution: true,
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
providers: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
id: "google",
|
||||
defaultModel: "gemini-3.1-flash-image-preview",
|
||||
authEnvVars: ["GEMINI_API_KEY", "GOOGLE_API_KEY"],
|
||||
models: expect.arrayContaining([
|
||||
"gemini-3.1-flash-image-preview",
|
||||
"gemini-3-pro-image-preview",
|
||||
]),
|
||||
capabilities: expect.objectContaining({
|
||||
edit: expect.objectContaining({
|
||||
enabled: true,
|
||||
maxInputImages: 5,
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
expect.objectContaining({
|
||||
id: "openai",
|
||||
authEnvVars: ["OPENAI_API_KEY"],
|
||||
}),
|
||||
]),
|
||||
},
|
||||
});
|
||||
expect(openaiProvider.authEnvVars).toEqual(["OPENAI_API_KEY"]);
|
||||
});
|
||||
|
||||
it("skips auth hints for prototype-like provider ids", async () => {
|
||||
@@ -1456,15 +1474,15 @@ describe("createImageGenerateTool", () => {
|
||||
);
|
||||
|
||||
const result = await tool.execute("call-list-proto", { action: "list" });
|
||||
const text = resultText(result);
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
|
||||
expect(text).toContain("__proto__ (default proto-v1)");
|
||||
expect(text).not.toContain("auth: set");
|
||||
const details = resultDetails(result);
|
||||
const providers = details.providers as Array<Record<string, unknown>>;
|
||||
expect(providers).toHaveLength(1);
|
||||
expect(providers[0]?.id).toBe("__proto__");
|
||||
expect(providers[0]?.authEnvVars).toEqual([]);
|
||||
expect(result).toMatchObject({
|
||||
details: {
|
||||
providers: [expect.objectContaining({ id: "__proto__", authEnvVars: [] })],
|
||||
},
|
||||
});
|
||||
});
|
||||
|
||||
it("rejects provider-specific edit limits before runtime", async () => {
|
||||
@@ -1554,9 +1572,13 @@ describe("createImageGenerateTool", () => {
|
||||
image: "./fixtures/a.png",
|
||||
aspectRatio: "16:9",
|
||||
});
|
||||
const text = resultText(result);
|
||||
const text = (result.content?.[0] as { text: string } | undefined)?.text ?? "";
|
||||
|
||||
expect(mockCallArg(generateImage, 0, "generateImage").aspectRatio).toBe("16:9");
|
||||
expect(generateImage).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
aspectRatio: "16:9",
|
||||
}),
|
||||
);
|
||||
expect(text).toContain(
|
||||
"Warning: Ignored unsupported overrides for fal/fal-ai/flux/dev: aspectRatio=16:9.",
|
||||
);
|
||||
|
||||
@@ -81,32 +81,6 @@ describe("sessions_spawn tool", () => {
|
||||
return property;
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function expectDetailFields(details: unknown, expected: Record<string, unknown>) {
|
||||
const record = requireRecord(details, "result details");
|
||||
for (const [key, value] of Object.entries(expected)) {
|
||||
expect(record[key]).toBe(value);
|
||||
}
|
||||
}
|
||||
|
||||
function mockCallArg(mock: unknown, callIndex: number, argIndex: number, label: string) {
|
||||
const calls = (mock as { mock?: { calls?: unknown[][] } }).mock?.calls;
|
||||
if (!Array.isArray(calls)) {
|
||||
throw new Error(`expected ${label} mock calls`);
|
||||
}
|
||||
const call = calls[callIndex];
|
||||
if (!call) {
|
||||
throw new Error(`expected ${label} call ${callIndex + 1}`);
|
||||
}
|
||||
return requireRecord(call[argIndex], `${label} call ${callIndex + 1} arg ${argIndex + 1}`);
|
||||
}
|
||||
|
||||
it("hides ACP runtime affordances when no ACP backend is loaded", () => {
|
||||
const tool = createSessionsSpawnTool();
|
||||
const schema = tool.parameters as {
|
||||
@@ -181,7 +155,10 @@ describe("sessions_spawn tool", () => {
|
||||
agentId: "codex",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { status: "error", role: "codex" });
|
||||
expect(result.details).toMatchObject({
|
||||
status: "error",
|
||||
role: "codex",
|
||||
});
|
||||
expect(JSON.stringify(result.details)).toContain("no ACP runtime backend is loaded");
|
||||
expect(hoisted.spawnAcpDirectMock).not.toHaveBeenCalled();
|
||||
expect(hoisted.spawnSubagentDirectMock).not.toHaveBeenCalled();
|
||||
@@ -291,23 +268,27 @@ describe("sessions_spawn tool", () => {
|
||||
cleanup: "keep",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, {
|
||||
expect(result.details).toMatchObject({
|
||||
status: "accepted",
|
||||
childSessionKey: "agent:main:subagent:1",
|
||||
runId: "run-subagent",
|
||||
});
|
||||
expect(result.details).not.toHaveProperty("role");
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("build feature");
|
||||
expect(spawnArgs.agentId).toBe("main");
|
||||
expect(spawnArgs.model).toBe("anthropic/claude-sonnet-4-6");
|
||||
expect(spawnArgs.thinking).toBe("medium");
|
||||
expect(spawnArgs.runTimeoutSeconds).toBe(5);
|
||||
expect(spawnArgs.thread).toBe(true);
|
||||
expect(spawnArgs.mode).toBe("session");
|
||||
expect(spawnArgs.cleanup).toBe("keep");
|
||||
const spawnContext = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 1, "spawnSubagentDirect");
|
||||
expect(spawnContext.agentSessionKey).toBe("agent:main:main");
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "build feature",
|
||||
agentId: "main",
|
||||
model: "anthropic/claude-sonnet-4-6",
|
||||
thinking: "medium",
|
||||
runTimeoutSeconds: 5,
|
||||
thread: true,
|
||||
mode: "session",
|
||||
cleanup: "keep",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
agentSessionKey: "agent:main:main",
|
||||
}),
|
||||
);
|
||||
expect(hoisted.spawnAcpDirectMock).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
@@ -328,13 +309,17 @@ describe("sessions_spawn tool", () => {
|
||||
taskName: "review_subagents",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, {
|
||||
expect(result.details).toMatchObject({
|
||||
status: "accepted",
|
||||
childSessionKey: "agent:main:subagent:1",
|
||||
});
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("review subagent handling");
|
||||
expect(spawnArgs.taskName).toBe("review_subagents");
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "review subagent handling",
|
||||
taskName: "review_subagents",
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
});
|
||||
|
||||
it("rejects invalid taskName before spawning", async () => {
|
||||
@@ -347,7 +332,9 @@ describe("sessions_spawn tool", () => {
|
||||
taskName: "Bad-Name",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { status: "error" });
|
||||
expect(result.details).toMatchObject({
|
||||
status: "error",
|
||||
});
|
||||
expect(JSON.stringify(result.details)).toContain("Invalid taskName");
|
||||
expect(hoisted.spawnSubagentDirectMock).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -362,7 +349,9 @@ describe("sessions_spawn tool", () => {
|
||||
taskName,
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { status: "error" });
|
||||
expect(result.details).toMatchObject({
|
||||
status: "error",
|
||||
});
|
||||
expect(JSON.stringify(result.details)).toContain("Reserved subagent targets");
|
||||
expect(hoisted.spawnSubagentDirectMock).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -381,7 +370,10 @@ describe("sessions_spawn tool", () => {
|
||||
agentId: "reviewer",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { ...spawnResult, role: "reviewer" });
|
||||
expect(result.details).toMatchObject({
|
||||
...spawnResult,
|
||||
role: "reviewer",
|
||||
});
|
||||
});
|
||||
|
||||
it("does not add role to forwarded failures when agentId is absent", async () => {
|
||||
@@ -397,7 +389,10 @@ describe("sessions_spawn tool", () => {
|
||||
task: "build feature",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { status: "error", error: "spawn failed" });
|
||||
expect(result.details).toMatchObject({
|
||||
status: "error",
|
||||
error: "spawn failed",
|
||||
});
|
||||
expect(result.details).not.toHaveProperty("role");
|
||||
});
|
||||
|
||||
@@ -411,9 +406,13 @@ describe("sessions_spawn tool", () => {
|
||||
timeoutSeconds: 2,
|
||||
});
|
||||
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("do thing");
|
||||
expect(spawnArgs.runTimeoutSeconds).toBe(2);
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "do thing",
|
||||
runTimeoutSeconds: 2,
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
});
|
||||
|
||||
it("passes inherited workspaceDir from tool context, not from tool args", async () => {
|
||||
@@ -427,8 +426,12 @@ describe("sessions_spawn tool", () => {
|
||||
workspaceDir: "/tmp/attempted-override",
|
||||
});
|
||||
|
||||
const spawnContext = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 1, "spawnSubagentDirect");
|
||||
expect(spawnContext.workspaceDir).toBe("/parent/workspace");
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.any(Object),
|
||||
expect.objectContaining({
|
||||
workspaceDir: "/parent/workspace",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("passes lightContext through to subagent spawns", async () => {
|
||||
@@ -441,9 +444,13 @@ describe("sessions_spawn tool", () => {
|
||||
lightContext: true,
|
||||
});
|
||||
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("summarize this");
|
||||
expect(spawnArgs.lightContext).toBe(true);
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "summarize this",
|
||||
lightContext: true,
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
});
|
||||
|
||||
it('rejects lightContext when runtime is not "subagent"', async () => {
|
||||
@@ -485,21 +492,25 @@ describe("sessions_spawn tool", () => {
|
||||
streamTo: "parent",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, {
|
||||
expect(result.details).toMatchObject({
|
||||
status: "accepted",
|
||||
childSessionKey: "agent:codex:acp:1",
|
||||
runId: "run-acp",
|
||||
});
|
||||
const spawnArgs = mockCallArg(hoisted.spawnAcpDirectMock, 0, 0, "spawnAcpDirect");
|
||||
expect(spawnArgs.task).toBe("investigate the failing CI run");
|
||||
expect(spawnArgs.agentId).toBe("codex");
|
||||
expect(spawnArgs.cwd).toBe("/workspace");
|
||||
expect(spawnArgs.runTimeoutSeconds).toBe(45);
|
||||
expect(spawnArgs.thread).toBe(true);
|
||||
expect(spawnArgs.mode).toBe("session");
|
||||
expect(spawnArgs.streamTo).toBe("parent");
|
||||
const spawnContext = mockCallArg(hoisted.spawnAcpDirectMock, 0, 1, "spawnAcpDirect");
|
||||
expect(spawnContext.agentSessionKey).toBe("agent:main:main");
|
||||
expect(hoisted.spawnAcpDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "investigate the failing CI run",
|
||||
agentId: "codex",
|
||||
cwd: "/workspace",
|
||||
runTimeoutSeconds: 45,
|
||||
thread: true,
|
||||
mode: "session",
|
||||
streamTo: "parent",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
agentSessionKey: "agent:main:main",
|
||||
}),
|
||||
);
|
||||
expect(hoisted.spawnSubagentDirectMock).not.toHaveBeenCalled();
|
||||
expect(hoisted.registerSubagentRunMock).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -517,10 +528,14 @@ describe("sessions_spawn tool", () => {
|
||||
model: "github-copilot/claude-sonnet-4.6",
|
||||
});
|
||||
|
||||
const spawnArgs = mockCallArg(hoisted.spawnAcpDirectMock, 0, 0, "spawnAcpDirect");
|
||||
expect(spawnArgs.task).toBe("investigate the failing CI run");
|
||||
expect(spawnArgs.agentId).toBe("codex");
|
||||
expect(spawnArgs.model).toBe("github-copilot/claude-sonnet-4.6");
|
||||
expect(hoisted.spawnAcpDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "investigate the failing CI run",
|
||||
agentId: "codex",
|
||||
model: "github-copilot/claude-sonnet-4.6",
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
});
|
||||
|
||||
it("adds requested role to forwarded ACP failures", async () => {
|
||||
@@ -540,7 +555,7 @@ describe("sessions_spawn tool", () => {
|
||||
agentId: "codex",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, {
|
||||
expect(result.details).toMatchObject({
|
||||
status: "forbidden",
|
||||
error: "ACP disabled",
|
||||
errorCode: "acp_disabled",
|
||||
@@ -561,18 +576,25 @@ describe("sessions_spawn tool", () => {
|
||||
sandbox: "require",
|
||||
});
|
||||
|
||||
const spawnArgs = mockCallArg(hoisted.spawnAcpDirectMock, 0, 0, "spawnAcpDirect");
|
||||
expect(spawnArgs.task).toBe("investigate");
|
||||
expect(spawnArgs.sandbox).toBe("require");
|
||||
const spawnContext = mockCallArg(hoisted.spawnAcpDirectMock, 0, 1, "spawnAcpDirect");
|
||||
expect(spawnContext.agentSessionKey).toBe("agent:main:subagent:parent");
|
||||
const registration = mockCallArg(hoisted.registerSubagentRunMock, 0, 0, "registerSubagentRun");
|
||||
expect(registration.runId).toBe("run-acp");
|
||||
expect(registration.childSessionKey).toBe("agent:codex:acp:1");
|
||||
expect(registration.requesterSessionKey).toBe("agent:main:subagent:parent");
|
||||
expect(registration.task).toBe("investigate");
|
||||
expect(registration.cleanup).toBe("keep");
|
||||
expect(registration.spawnMode).toBe("run");
|
||||
expect(hoisted.spawnAcpDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "investigate",
|
||||
sandbox: "require",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
agentSessionKey: "agent:main:subagent:parent",
|
||||
}),
|
||||
);
|
||||
expect(hoisted.registerSubagentRunMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
runId: "run-acp",
|
||||
childSessionKey: "agent:codex:acp:1",
|
||||
requesterSessionKey: "agent:main:subagent:parent",
|
||||
task: "investigate",
|
||||
cleanup: "keep",
|
||||
spawnMode: "run",
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("suppresses completion announces for inline ACP session delivery", async () => {
|
||||
@@ -600,14 +622,17 @@ describe("sessions_spawn tool", () => {
|
||||
mode: "session",
|
||||
});
|
||||
|
||||
const registration = mockCallArg(hoisted.registerSubagentRunMock, 0, 0, "registerSubagentRun");
|
||||
expect(registration.runId).toBe("run-acp");
|
||||
expect(registration.childSessionKey).toBe("agent:codex:acp:1");
|
||||
expect(registration.requesterSessionKey).toBe("agent:main:main");
|
||||
expect(registration.task).toBe("investigate");
|
||||
expect(registration.cleanup).toBe("keep");
|
||||
expect(registration.spawnMode).toBe("session");
|
||||
expect(registration.expectsCompletionMessage).toBe(false);
|
||||
expect(hoisted.registerSubagentRunMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
runId: "run-acp",
|
||||
childSessionKey: "agent:codex:acp:1",
|
||||
requesterSessionKey: "agent:main:main",
|
||||
task: "investigate",
|
||||
cleanup: "keep",
|
||||
spawnMode: "session",
|
||||
expectsCompletionMessage: false,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("rejects ACP runtime calls from sandboxed requester sessions", async () => {
|
||||
@@ -623,7 +648,10 @@ describe("sessions_spawn tool", () => {
|
||||
agentId: "codex",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { status: "error", role: "codex" });
|
||||
expect(result.details).toMatchObject({
|
||||
status: "error",
|
||||
role: "codex",
|
||||
});
|
||||
expect(JSON.stringify(result.details)).toContain("sandboxed sessions");
|
||||
expect(hoisted.spawnAcpDirectMock).not.toHaveBeenCalled();
|
||||
});
|
||||
@@ -641,10 +669,14 @@ describe("sessions_spawn tool", () => {
|
||||
resumeSessionId: "7f4a78e0-f6be-43fe-855c-c1c4fd229bc4",
|
||||
});
|
||||
|
||||
const spawnArgs = mockCallArg(hoisted.spawnAcpDirectMock, 0, 0, "spawnAcpDirect");
|
||||
expect(spawnArgs.task).toBe("resume prior work");
|
||||
expect(spawnArgs.agentId).toBe("codex");
|
||||
expect(spawnArgs.resumeSessionId).toBe("7f4a78e0-f6be-43fe-855c-c1c4fd229bc4");
|
||||
expect(hoisted.spawnAcpDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "resume prior work",
|
||||
agentId: "codex",
|
||||
resumeSessionId: "7f4a78e0-f6be-43fe-855c-c1c4fd229bc4",
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
});
|
||||
|
||||
it("ignores ACP-only fields for subagent spawns", async () => {
|
||||
@@ -659,15 +691,19 @@ describe("sessions_spawn tool", () => {
|
||||
streamTo: "parent",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, {
|
||||
expect(result.details).toMatchObject({
|
||||
status: "accepted",
|
||||
childSessionKey: "agent:main:subagent:1",
|
||||
runId: "run-subagent",
|
||||
});
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("resume prior work");
|
||||
const spawnContext = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 1, "spawnSubagentDirect");
|
||||
expect(spawnContext.agentSessionKey).toBe("agent:main:main");
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "resume prior work",
|
||||
}),
|
||||
expect.objectContaining({
|
||||
agentSessionKey: "agent:main:main",
|
||||
}),
|
||||
);
|
||||
expect(hoisted.spawnSubagentDirectMock.mock.calls[0]?.[0]).not.toHaveProperty(
|
||||
"resumeSessionId",
|
||||
);
|
||||
@@ -691,7 +727,9 @@ describe("sessions_spawn tool", () => {
|
||||
attachments: [{ name: "a.txt", content: "hello", encoding: "utf8" }],
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, { status: "error" });
|
||||
expect(result.details).toMatchObject({
|
||||
status: "error",
|
||||
});
|
||||
const details = result.details as { error?: string };
|
||||
expect(details.error).toContain("attachments are currently unsupported for runtime=acp");
|
||||
expect(hoisted.spawnAcpDirectMock).not.toHaveBeenCalled();
|
||||
@@ -709,14 +747,18 @@ describe("sessions_spawn tool", () => {
|
||||
streamTo: "parent",
|
||||
});
|
||||
|
||||
expectDetailFields(result.details, {
|
||||
expect(result.details).toMatchObject({
|
||||
status: "accepted",
|
||||
childSessionKey: "agent:main:subagent:1",
|
||||
runId: "run-subagent",
|
||||
});
|
||||
expect(hoisted.spawnAcpDirectMock).not.toHaveBeenCalled();
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("analyze file");
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "analyze file",
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
expect(hoisted.spawnSubagentDirectMock.mock.calls[0]?.[0]).not.toHaveProperty(
|
||||
"resumeSessionId",
|
||||
);
|
||||
@@ -733,9 +775,13 @@ describe("sessions_spawn tool", () => {
|
||||
model: "default",
|
||||
});
|
||||
|
||||
const spawnArgs = mockCallArg(hoisted.spawnSubagentDirectMock, 0, 0, "spawnSubagentDirect");
|
||||
expect(spawnArgs.task).toBe("analyze file");
|
||||
expect(spawnArgs.model).toBeUndefined();
|
||||
expect(hoisted.spawnSubagentDirectMock).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
task: "analyze file",
|
||||
model: undefined,
|
||||
}),
|
||||
expect.any(Object),
|
||||
);
|
||||
});
|
||||
|
||||
it("keeps attachment content schema unconstrained for llama.cpp grammar safety", () => {
|
||||
|
||||
@@ -3652,45 +3652,8 @@ describe("runAgentTurnWithFallback", () => {
|
||||
providerOverride: "anthropic",
|
||||
modelOverride: "claude-sonnet",
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: "anthropic",
|
||||
modelOverrideFallbackOriginModel: "claude-opus",
|
||||
authProfileOverride: "anthropic:openclaw",
|
||||
authProfileOverrideSource: "user",
|
||||
});
|
||||
});
|
||||
|
||||
it("preserves original auto-fallback origin across chained fallbacks", async () => {
|
||||
const applyFallbackCandidateSelectionToEntry =
|
||||
await getApplyFallbackCandidateSelectionToEntry();
|
||||
const entry = {
|
||||
sessionId: "session",
|
||||
updatedAt: 1,
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "fallback-b",
|
||||
modelOverrideSource: "auto" as const,
|
||||
modelOverrideFallbackOriginProvider: "anthropic",
|
||||
modelOverrideFallbackOriginModel: "claude-opus",
|
||||
} as SessionEntry;
|
||||
|
||||
const { updated } = applyFallbackCandidateSelectionToEntry({
|
||||
entry,
|
||||
run: {
|
||||
provider: "openrouter",
|
||||
model: "fallback-b",
|
||||
} as FollowupRun["run"],
|
||||
provider: "openrouter",
|
||||
model: "fallback-c",
|
||||
now: 123,
|
||||
});
|
||||
|
||||
expect(updated).toBe(true);
|
||||
expect(entry).toMatchObject({
|
||||
updatedAt: 123,
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "fallback-c",
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: "anthropic",
|
||||
modelOverrideFallbackOriginModel: "claude-opus",
|
||||
});
|
||||
});
|
||||
});
|
||||
|
||||
@@ -130,8 +130,6 @@ type FallbackSelectionState = Pick<
|
||||
| "providerOverride"
|
||||
| "modelOverride"
|
||||
| "modelOverrideSource"
|
||||
| "modelOverrideFallbackOriginProvider"
|
||||
| "modelOverrideFallbackOriginModel"
|
||||
| "authProfileOverride"
|
||||
| "authProfileOverrideSource"
|
||||
| "authProfileOverrideCompactionCount"
|
||||
@@ -141,8 +139,6 @@ const FALLBACK_SELECTION_STATE_KEYS = [
|
||||
"providerOverride",
|
||||
"modelOverride",
|
||||
"modelOverrideSource",
|
||||
"modelOverrideFallbackOriginProvider",
|
||||
"modelOverrideFallbackOriginModel",
|
||||
"authProfileOverride",
|
||||
"authProfileOverrideSource",
|
||||
"authProfileOverrideCompactionCount",
|
||||
@@ -172,20 +168,6 @@ function setFallbackSelectionStateField(
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
case "modelOverrideFallbackOriginProvider":
|
||||
if (entry.modelOverrideFallbackOriginProvider !== value) {
|
||||
entry.modelOverrideFallbackOriginProvider =
|
||||
value as SessionEntry["modelOverrideFallbackOriginProvider"];
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
case "modelOverrideFallbackOriginModel":
|
||||
if (entry.modelOverrideFallbackOriginModel !== value) {
|
||||
entry.modelOverrideFallbackOriginModel =
|
||||
value as SessionEntry["modelOverrideFallbackOriginModel"];
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
case "authProfileOverride":
|
||||
if (entry.authProfileOverride !== value) {
|
||||
entry.authProfileOverride = value as SessionEntry["authProfileOverride"];
|
||||
@@ -214,8 +196,6 @@ function snapshotFallbackSelectionState(entry: SessionEntry): FallbackSelectionS
|
||||
providerOverride: entry.providerOverride,
|
||||
modelOverride: entry.modelOverride,
|
||||
modelOverrideSource: entry.modelOverrideSource,
|
||||
modelOverrideFallbackOriginProvider: entry.modelOverrideFallbackOriginProvider,
|
||||
modelOverrideFallbackOriginModel: entry.modelOverrideFallbackOriginModel,
|
||||
authProfileOverride: entry.authProfileOverride,
|
||||
authProfileOverrideSource: entry.authProfileOverrideSource,
|
||||
authProfileOverrideCompactionCount: entry.authProfileOverrideCompactionCount,
|
||||
@@ -225,8 +205,6 @@ function snapshotFallbackSelectionState(entry: SessionEntry): FallbackSelectionS
|
||||
function buildFallbackSelectionState(params: {
|
||||
provider: string;
|
||||
model: string;
|
||||
originProvider: string;
|
||||
originModel: string;
|
||||
authProfileId?: string;
|
||||
authProfileIdSource?: "auto" | "user";
|
||||
}): FallbackSelectionState {
|
||||
@@ -234,32 +212,12 @@ function buildFallbackSelectionState(params: {
|
||||
providerOverride: params.provider,
|
||||
modelOverride: params.model,
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: params.originProvider,
|
||||
modelOverrideFallbackOriginModel: params.originModel,
|
||||
authProfileOverride: params.authProfileId,
|
||||
authProfileOverrideSource: params.authProfileId ? params.authProfileIdSource : undefined,
|
||||
authProfileOverrideCompactionCount: undefined,
|
||||
};
|
||||
}
|
||||
|
||||
function resolveFallbackSelectionOrigin(params: { entry: SessionEntry; run: FollowupRun["run"] }): {
|
||||
provider: string;
|
||||
model: string;
|
||||
} {
|
||||
if (params.entry.modelOverrideSource === "auto") {
|
||||
const persistedOriginProvider = normalizeOptionalString(
|
||||
params.entry.modelOverrideFallbackOriginProvider,
|
||||
);
|
||||
const persistedOriginModel = normalizeOptionalString(
|
||||
params.entry.modelOverrideFallbackOriginModel,
|
||||
);
|
||||
if (persistedOriginProvider && persistedOriginModel) {
|
||||
return { provider: persistedOriginProvider, model: persistedOriginModel };
|
||||
}
|
||||
}
|
||||
return { provider: params.run.provider, model: params.run.model };
|
||||
}
|
||||
|
||||
export function applyFallbackCandidateSelectionToEntry(params: {
|
||||
entry: SessionEntry;
|
||||
run: FollowupRun["run"];
|
||||
@@ -271,12 +229,9 @@ export function applyFallbackCandidateSelectionToEntry(params: {
|
||||
return { updated: false };
|
||||
}
|
||||
const scopedAuthProfile = resolveRunAuthProfile(params.run, params.provider);
|
||||
const origin = resolveFallbackSelectionOrigin({ entry: params.entry, run: params.run });
|
||||
const nextState = buildFallbackSelectionState({
|
||||
provider: params.provider,
|
||||
model: params.model,
|
||||
originProvider: origin.provider,
|
||||
originModel: origin.model,
|
||||
authProfileId: scopedAuthProfile.authProfileId,
|
||||
authProfileIdSource: scopedAuthProfile.authProfileIdSource,
|
||||
});
|
||||
|
||||
@@ -166,8 +166,6 @@ export async function resolveReplyDirectives(params: {
|
||||
commandAuthorized: boolean;
|
||||
defaultProvider: string;
|
||||
defaultModel: string;
|
||||
primaryProvider?: string;
|
||||
primaryModel?: string;
|
||||
aliasIndex: ModelAliasIndex;
|
||||
provider: string;
|
||||
model: string;
|
||||
@@ -196,8 +194,6 @@ export async function resolveReplyDirectives(params: {
|
||||
commandAuthorized,
|
||||
defaultProvider,
|
||||
defaultModel,
|
||||
primaryProvider,
|
||||
primaryModel,
|
||||
provider: initialProvider,
|
||||
model: initialModel,
|
||||
hasResolvedHeartbeatModelOverride,
|
||||
@@ -530,13 +526,10 @@ export async function resolveReplyDirectives(params: {
|
||||
storePath,
|
||||
defaultProvider,
|
||||
defaultModel,
|
||||
primaryProvider,
|
||||
primaryModel,
|
||||
provider,
|
||||
model,
|
||||
hasModelDirective: directives.hasModelDirective,
|
||||
hasResolvedHeartbeatModelOverride,
|
||||
isHeartbeat: opts?.isHeartbeat === true,
|
||||
});
|
||||
provider = modelState.provider;
|
||||
model = modelState.model;
|
||||
|
||||
@@ -43,10 +43,7 @@ import { hasInboundMedia } from "./inbound-media.js";
|
||||
import { emitPreAgentMessageHooks } from "./message-preprocess-hooks.js";
|
||||
import { createFastTestModelSelectionState } from "./model-selection.js";
|
||||
import { initSessionState } from "./session.js";
|
||||
import {
|
||||
isStaleHeartbeatAutoFallbackOverride,
|
||||
resolveStoredModelOverride,
|
||||
} from "./stored-model-override.js";
|
||||
import { resolveStoredModelOverride } from "./stored-model-override.js";
|
||||
import { createTypingController } from "./typing.js";
|
||||
|
||||
type ResetCommandAction = "new" | "reset";
|
||||
@@ -435,16 +432,6 @@ export async function getReplyFromConfig(
|
||||
parentSessionKey: sessionCtx.ModelParentSessionKey ?? sessionCtx.ParentSessionKey,
|
||||
})
|
||||
: null;
|
||||
const resolvedChannelModelOverride =
|
||||
channelModelOverride && !hasResolvedHeartbeatModelOverride
|
||||
? resolveModelRefFromString({
|
||||
raw: channelModelOverride.model,
|
||||
defaultProvider,
|
||||
aliasIndex,
|
||||
})
|
||||
: null;
|
||||
const primaryProvider = resolvedChannelModelOverride?.ref.provider ?? defaultProvider;
|
||||
const primaryModel = resolvedChannelModelOverride?.ref.model ?? defaultModel;
|
||||
const hasSessionModelOverride = Boolean(
|
||||
normalizeOptionalString(sessionEntry.modelOverride) ||
|
||||
normalizeOptionalString(sessionEntry.providerOverride),
|
||||
@@ -459,33 +446,20 @@ export async function getReplyFromConfig(
|
||||
sessionCtx.ParentSessionKey,
|
||||
defaultProvider,
|
||||
});
|
||||
const staleHeartbeatAutoFallbackOverride = isStaleHeartbeatAutoFallbackOverride({
|
||||
isHeartbeat: opts?.isHeartbeat === true,
|
||||
hasResolvedHeartbeatModelOverride,
|
||||
sessionEntry,
|
||||
storedOverride: storedModelOverride,
|
||||
defaultProvider,
|
||||
defaultModel,
|
||||
primaryProvider,
|
||||
primaryModel,
|
||||
});
|
||||
if (
|
||||
storedModelOverride?.model &&
|
||||
!hasResolvedHeartbeatModelOverride &&
|
||||
!staleHeartbeatAutoFallbackOverride
|
||||
) {
|
||||
if (storedModelOverride?.model && !hasResolvedHeartbeatModelOverride) {
|
||||
provider = storedModelOverride.provider ?? defaultProvider;
|
||||
model = storedModelOverride.model;
|
||||
}
|
||||
const hasEffectiveSessionModelOverride =
|
||||
hasSessionModelOverride && !staleHeartbeatAutoFallbackOverride;
|
||||
if (
|
||||
!hasResolvedHeartbeatModelOverride &&
|
||||
!hasEffectiveSessionModelOverride &&
|
||||
resolvedChannelModelOverride
|
||||
) {
|
||||
provider = resolvedChannelModelOverride.ref.provider;
|
||||
model = resolvedChannelModelOverride.ref.model;
|
||||
if (!hasResolvedHeartbeatModelOverride && !hasSessionModelOverride && channelModelOverride) {
|
||||
const resolved = resolveModelRefFromString({
|
||||
raw: channelModelOverride.model,
|
||||
defaultProvider,
|
||||
aliasIndex,
|
||||
});
|
||||
if (resolved) {
|
||||
provider = resolved.ref.provider;
|
||||
model = resolved.ref.model;
|
||||
}
|
||||
}
|
||||
|
||||
if (
|
||||
@@ -586,8 +560,6 @@ export async function getReplyFromConfig(
|
||||
commandAuthorized,
|
||||
defaultProvider,
|
||||
defaultModel,
|
||||
primaryProvider,
|
||||
primaryModel,
|
||||
aliasIndex,
|
||||
provider,
|
||||
model,
|
||||
|
||||
@@ -852,27 +852,12 @@ describe("createModelSelectionState auto-failover overrides", () => {
|
||||
providerOverride: string;
|
||||
modelOverride: string;
|
||||
modelOverrideSource: "auto" | "user" | undefined;
|
||||
modelOverrideFallbackOriginProvider?: string;
|
||||
modelOverrideFallbackOriginModel?: string;
|
||||
fallbackNoticeSelectedModel?: string;
|
||||
authProfileOverride?: string;
|
||||
authProfileOverrideSource?: "auto" | "user";
|
||||
provider?: string;
|
||||
model?: string;
|
||||
primaryProvider?: string;
|
||||
primaryModel?: string;
|
||||
isHeartbeat?: boolean;
|
||||
}) {
|
||||
const cfg = {} as OpenClawConfig;
|
||||
const sessionEntry = makeEntry({
|
||||
providerOverride: params.providerOverride,
|
||||
modelOverride: params.modelOverride,
|
||||
modelOverrideSource: params.modelOverrideSource,
|
||||
modelOverrideFallbackOriginProvider: params.modelOverrideFallbackOriginProvider,
|
||||
modelOverrideFallbackOriginModel: params.modelOverrideFallbackOriginModel,
|
||||
fallbackNoticeSelectedModel: params.fallbackNoticeSelectedModel,
|
||||
authProfileOverride: params.authProfileOverride,
|
||||
authProfileOverrideSource: params.authProfileOverrideSource,
|
||||
});
|
||||
const sessionStore = { [sessionKey]: sessionEntry };
|
||||
const state = await createModelSelectionState({
|
||||
@@ -883,12 +868,9 @@ describe("createModelSelectionState auto-failover overrides", () => {
|
||||
sessionKey,
|
||||
defaultProvider,
|
||||
defaultModel,
|
||||
primaryProvider: params.primaryProvider,
|
||||
primaryModel: params.primaryModel,
|
||||
provider: params.provider ?? defaultProvider,
|
||||
model: params.model ?? defaultModel,
|
||||
provider: defaultProvider,
|
||||
model: defaultModel,
|
||||
hasModelDirective: false,
|
||||
isHeartbeat: params.isHeartbeat,
|
||||
});
|
||||
return { state, sessionEntry, sessionStore };
|
||||
}
|
||||
@@ -972,137 +954,6 @@ describe("createModelSelectionState auto-failover overrides", () => {
|
||||
expect(state.resetModelOverride).toBe(false);
|
||||
});
|
||||
|
||||
it("clears stale heartbeat auto-failover override when the fallback origin changed", async () => {
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "minimax/minimax-m2.7",
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: "openai-codex",
|
||||
modelOverrideFallbackOriginModel: "gpt-5.3",
|
||||
provider: "openrouter",
|
||||
model: "minimax/minimax-m2.7",
|
||||
isHeartbeat: true,
|
||||
});
|
||||
|
||||
expect(state.provider).toBe(defaultProvider);
|
||||
expect(state.model).toBe(defaultModel);
|
||||
expect(state.resetModelOverride).toBe(true);
|
||||
expect(state.resetModelOverrideRef).toBe("openrouter/minimax/minimax-m2.7");
|
||||
expect(sessionStore[sessionKey]?.providerOverride).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverride).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverrideSource).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverrideFallbackOriginProvider).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverrideFallbackOriginModel).toBeUndefined();
|
||||
});
|
||||
|
||||
it("preserves user auth profile when clearing a stale heartbeat auto-failover override", async () => {
|
||||
authProfileStoreMock.store = {
|
||||
version: 1,
|
||||
profiles: {
|
||||
"mac-studio:local": {
|
||||
type: "api_key",
|
||||
provider: defaultProvider,
|
||||
key: "test-key",
|
||||
},
|
||||
},
|
||||
};
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "minimax/minimax-m2.7",
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: "openai-codex",
|
||||
modelOverrideFallbackOriginModel: "gpt-5.3",
|
||||
authProfileOverride: "mac-studio:local",
|
||||
authProfileOverrideSource: "user",
|
||||
provider: "openrouter",
|
||||
model: "minimax/minimax-m2.7",
|
||||
isHeartbeat: true,
|
||||
});
|
||||
|
||||
expect(state.provider).toBe(defaultProvider);
|
||||
expect(state.model).toBe(defaultModel);
|
||||
expect(state.resetModelOverride).toBe(true);
|
||||
expect(sessionStore[sessionKey]?.providerOverride).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverride).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.authProfileOverride).toBe("mac-studio:local");
|
||||
expect(sessionStore[sessionKey]?.authProfileOverrideSource).toBe("user");
|
||||
});
|
||||
|
||||
it("keeps heartbeat auto-failover override when the fallback origin still matches default", async () => {
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "minimax/minimax-m2.7",
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: defaultProvider,
|
||||
modelOverrideFallbackOriginModel: defaultModel,
|
||||
provider: "openrouter",
|
||||
model: "minimax/minimax-m2.7",
|
||||
isHeartbeat: true,
|
||||
});
|
||||
|
||||
expect(state.provider).toBe("openrouter");
|
||||
expect(state.model).toBe("minimax/minimax-m2.7");
|
||||
expect(state.resetModelOverride).toBe(false);
|
||||
expect(sessionStore[sessionKey]?.providerOverride).toBe("openrouter");
|
||||
expect(sessionStore[sessionKey]?.modelOverride).toBe("minimax/minimax-m2.7");
|
||||
});
|
||||
|
||||
it("keeps heartbeat auto-failover override when the origin matches the channel primary", async () => {
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "minimax/minimax-m2.7",
|
||||
modelOverrideSource: "auto",
|
||||
modelOverrideFallbackOriginProvider: "openai",
|
||||
modelOverrideFallbackOriginModel: "gpt-4o",
|
||||
primaryProvider: "openai",
|
||||
primaryModel: "gpt-4o",
|
||||
provider: "openrouter",
|
||||
model: "minimax/minimax-m2.7",
|
||||
isHeartbeat: true,
|
||||
});
|
||||
|
||||
expect(state.provider).toBe("openrouter");
|
||||
expect(state.model).toBe("minimax/minimax-m2.7");
|
||||
expect(state.resetModelOverride).toBe(false);
|
||||
expect(sessionStore[sessionKey]?.providerOverride).toBe("openrouter");
|
||||
expect(sessionStore[sessionKey]?.modelOverride).toBe("minimax/minimax-m2.7");
|
||||
});
|
||||
|
||||
it("clears legacy heartbeat auto-failover override when no origin metadata exists", async () => {
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "minimax/minimax-m2.7",
|
||||
modelOverrideSource: "auto",
|
||||
provider: "openrouter",
|
||||
model: "minimax/minimax-m2.7",
|
||||
isHeartbeat: true,
|
||||
});
|
||||
|
||||
expect(state.provider).toBe(defaultProvider);
|
||||
expect(state.model).toBe(defaultModel);
|
||||
expect(state.resetModelOverride).toBe(true);
|
||||
expect(sessionStore[sessionKey]?.providerOverride).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverride).toBeUndefined();
|
||||
expect(sessionStore[sessionKey]?.modelOverrideSource).toBeUndefined();
|
||||
});
|
||||
|
||||
it("uses fallback notice metadata for legacy heartbeat auto-failover overrides", async () => {
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
modelOverride: "minimax/minimax-m2.7",
|
||||
modelOverrideSource: "auto",
|
||||
fallbackNoticeSelectedModel: `${defaultProvider}/${defaultModel}`,
|
||||
provider: "openrouter",
|
||||
model: "minimax/minimax-m2.7",
|
||||
isHeartbeat: true,
|
||||
});
|
||||
|
||||
expect(state.provider).toBe("openrouter");
|
||||
expect(state.model).toBe("minimax/minimax-m2.7");
|
||||
expect(state.resetModelOverride).toBe(false);
|
||||
expect(sessionStore[sessionKey]?.modelOverrideSource).toBe("auto");
|
||||
});
|
||||
|
||||
it("preserves a user-selected override across turns", async () => {
|
||||
const { state, sessionStore } = await resolveStateWithOverride({
|
||||
providerOverride: "openrouter",
|
||||
|
||||
@@ -28,10 +28,7 @@ export {
|
||||
resolveModelDirectiveSelection,
|
||||
type ModelDirectiveSelection,
|
||||
} from "./model-selection-directive.js";
|
||||
import {
|
||||
isStaleHeartbeatAutoFallbackOverride,
|
||||
resolveStoredModelOverride,
|
||||
} from "./stored-model-override.js";
|
||||
import { resolveStoredModelOverride } from "./stored-model-override.js";
|
||||
|
||||
type ModelCatalog = ModelCatalogEntry[];
|
||||
|
||||
@@ -98,15 +95,12 @@ export async function createModelSelectionState(params: {
|
||||
storePath?: string;
|
||||
defaultProvider: string;
|
||||
defaultModel: string;
|
||||
primaryProvider?: string;
|
||||
primaryModel?: string;
|
||||
provider: string;
|
||||
model: string;
|
||||
hasModelDirective: boolean;
|
||||
/** True when heartbeat.model was explicitly resolved for this run.
|
||||
* In that case, skip session-stored overrides so the heartbeat selection wins. */
|
||||
hasResolvedHeartbeatModelOverride?: boolean;
|
||||
isHeartbeat?: boolean;
|
||||
}): Promise<ModelSelectionState> {
|
||||
const timingEnabled = shouldLogModelSelectionTiming();
|
||||
const startMs = timingEnabled ? Date.now() : 0;
|
||||
@@ -133,8 +127,6 @@ export async function createModelSelectionState(params: {
|
||||
|
||||
let provider = params.provider;
|
||||
let model = params.model;
|
||||
const primaryProvider = params.primaryProvider ?? defaultProvider;
|
||||
const primaryModel = params.primaryModel ?? defaultModel;
|
||||
|
||||
const hasAllowlist = agentCfg?.models && Object.keys(agentCfg.models).length > 0;
|
||||
const visibility = parseConfiguredModelVisibilityEntries({ cfg });
|
||||
@@ -166,19 +158,6 @@ export async function createModelSelectionState(params: {
|
||||
overrideProvider: sessionEntry?.providerOverride,
|
||||
overrideModel: sessionEntry?.modelOverride,
|
||||
});
|
||||
const directStoredModelOverride = directStoredOverride
|
||||
? { ...directStoredOverride, source: "session" as const }
|
||||
: null;
|
||||
const staleHeartbeatAutoFallbackOverride = isStaleHeartbeatAutoFallbackOverride({
|
||||
isHeartbeat: params.isHeartbeat,
|
||||
hasResolvedHeartbeatModelOverride: params.hasResolvedHeartbeatModelOverride,
|
||||
sessionEntry,
|
||||
storedOverride: directStoredModelOverride,
|
||||
defaultProvider,
|
||||
defaultModel,
|
||||
primaryProvider: params.primaryProvider,
|
||||
primaryModel: params.primaryModel,
|
||||
});
|
||||
|
||||
if (needsModelCatalog) {
|
||||
modelCatalog = await (await loadModelCatalogRuntime()).loadModelCatalog({ config: cfg });
|
||||
@@ -220,11 +199,10 @@ export async function createModelSelectionState(params: {
|
||||
directStoredOverride.model,
|
||||
);
|
||||
const key = modelKey(normalizedOverride.provider, normalizedOverride.model);
|
||||
if (staleHeartbeatAutoFallbackOverride || !visibilityPolicy.allowsKey(key)) {
|
||||
if (!visibilityPolicy.allowsKey(key)) {
|
||||
const { updated } = applyModelOverrideToSessionEntry({
|
||||
entry: sessionEntry,
|
||||
selection: { provider: primaryProvider, model: primaryModel, isDefault: true },
|
||||
preserveAuthProfileOverride: staleHeartbeatAutoFallbackOverride,
|
||||
selection: { provider: defaultProvider, model: defaultModel, isDefault: true },
|
||||
});
|
||||
if (updated) {
|
||||
sessionStore[sessionKey] = sessionEntry;
|
||||
@@ -242,23 +220,6 @@ export async function createModelSelectionState(params: {
|
||||
}
|
||||
}
|
||||
}
|
||||
if (staleHeartbeatAutoFallbackOverride) {
|
||||
const normalizedCurrentSelection = normalizeModelRef(provider, model);
|
||||
const currentSelectionKey = modelKey(
|
||||
normalizedCurrentSelection.provider,
|
||||
normalizedCurrentSelection.model,
|
||||
);
|
||||
const normalizedDirectOverride = directStoredOverride
|
||||
? normalizeModelRef(directStoredOverride.provider, directStoredOverride.model)
|
||||
: null;
|
||||
const directStoredOverrideKey = normalizedDirectOverride
|
||||
? modelKey(normalizedDirectOverride.provider, normalizedDirectOverride.model)
|
||||
: undefined;
|
||||
if (currentSelectionKey === directStoredOverrideKey) {
|
||||
provider = primaryProvider;
|
||||
model = primaryModel;
|
||||
}
|
||||
}
|
||||
|
||||
const storedOverride = resolveStoredModelOverride({
|
||||
sessionEntry,
|
||||
@@ -268,12 +229,9 @@ export async function createModelSelectionState(params: {
|
||||
defaultProvider,
|
||||
});
|
||||
// Skip stored session model override only when an explicit heartbeat.model
|
||||
// was resolved. Heartbeats without heartbeat.model still inherit normal
|
||||
// overrides unless a direct auto fallback override is stale for the current
|
||||
// configured default.
|
||||
const skipStoredOverride =
|
||||
params.hasResolvedHeartbeatModelOverride === true ||
|
||||
(staleHeartbeatAutoFallbackOverride && storedOverride?.source === "session");
|
||||
// was resolved. Heartbeat runs without heartbeat.model should still inherit
|
||||
// the regular session/parent model override behavior.
|
||||
const skipStoredOverride = params.hasResolvedHeartbeatModelOverride === true;
|
||||
|
||||
if (storedOverride?.model && !skipStoredOverride) {
|
||||
const normalizedStoredOverride = normalizeModelRef(
|
||||
|
||||
@@ -1,8 +1,4 @@
|
||||
import {
|
||||
modelKey,
|
||||
normalizeModelRef,
|
||||
resolvePersistedOverrideModelRef,
|
||||
} from "../../agents/model-selection.js";
|
||||
import { resolvePersistedOverrideModelRef } from "../../agents/model-selection.js";
|
||||
import { resolveSessionParentSessionKey } from "../../channels/plugins/session-conversation.js";
|
||||
import type { SessionEntry } from "../../config/sessions/types.js";
|
||||
import { normalizeOptionalString } from "../../shared/string-coerce.js";
|
||||
@@ -61,70 +57,3 @@ export function resolveStoredModelOverride(params: {
|
||||
}
|
||||
return { ...parentOverride, source: "parent" };
|
||||
}
|
||||
|
||||
function resolveModelRefKey(params: {
|
||||
defaultProvider: string;
|
||||
overrideProvider?: string;
|
||||
overrideModel?: string;
|
||||
}): string | null {
|
||||
const ref = resolvePersistedOverrideModelRef(params);
|
||||
if (!ref) {
|
||||
return null;
|
||||
}
|
||||
const normalized = normalizeModelRef(ref.provider, ref.model);
|
||||
return modelKey(normalized.provider, normalized.model);
|
||||
}
|
||||
|
||||
export function isStaleHeartbeatAutoFallbackOverride(params: {
|
||||
isHeartbeat?: boolean;
|
||||
hasResolvedHeartbeatModelOverride?: boolean;
|
||||
sessionEntry?: SessionEntry;
|
||||
storedOverride?: StoredModelOverride | null;
|
||||
defaultProvider: string;
|
||||
defaultModel: string;
|
||||
primaryProvider?: string;
|
||||
primaryModel?: string;
|
||||
}): boolean {
|
||||
if (params.isHeartbeat !== true || params.hasResolvedHeartbeatModelOverride === true) {
|
||||
return false;
|
||||
}
|
||||
if (params.storedOverride?.source !== "session") {
|
||||
return false;
|
||||
}
|
||||
if (params.sessionEntry?.modelOverrideSource !== "auto") {
|
||||
return false;
|
||||
}
|
||||
|
||||
const primaryKey = resolveModelRefKey({
|
||||
defaultProvider: params.defaultProvider,
|
||||
overrideProvider: params.primaryProvider ?? params.defaultProvider,
|
||||
overrideModel: params.primaryModel ?? params.defaultModel,
|
||||
});
|
||||
if (!primaryKey) {
|
||||
return false;
|
||||
}
|
||||
|
||||
const originKey = resolveModelRefKey({
|
||||
defaultProvider: params.defaultProvider,
|
||||
overrideProvider: params.sessionEntry.modelOverrideFallbackOriginProvider,
|
||||
overrideModel: params.sessionEntry.modelOverrideFallbackOriginModel,
|
||||
});
|
||||
if (originKey) {
|
||||
return originKey !== primaryKey;
|
||||
}
|
||||
|
||||
const noticeSelectedKey = resolveModelRefKey({
|
||||
defaultProvider: params.defaultProvider,
|
||||
overrideModel: normalizeOptionalString(params.sessionEntry.fallbackNoticeSelectedModel),
|
||||
});
|
||||
if (noticeSelectedKey) {
|
||||
return noticeSelectedKey !== primaryKey;
|
||||
}
|
||||
|
||||
const storedOverrideKey = resolveModelRefKey({
|
||||
defaultProvider: params.defaultProvider,
|
||||
overrideProvider: params.storedOverride.provider,
|
||||
overrideModel: params.storedOverride.model,
|
||||
});
|
||||
return storedOverrideKey !== null && storedOverrideKey !== primaryKey;
|
||||
}
|
||||
|
||||
@@ -63,70 +63,6 @@ const localSnapshot = {
|
||||
file: { version: 1, agents: {} } as ExecApprovalsFile,
|
||||
};
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
if (!value || typeof value !== "object" || Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function requireArray(value: unknown, label: string): unknown[] {
|
||||
if (!Array.isArray(value)) {
|
||||
throw new Error(`Expected ${label}`);
|
||||
}
|
||||
return value;
|
||||
}
|
||||
|
||||
function expectFields(
|
||||
value: unknown,
|
||||
label: string,
|
||||
fields: Record<string, unknown>,
|
||||
): Record<string, unknown> {
|
||||
const record = requireRecord(value, label);
|
||||
for (const [key, expected] of Object.entries(fields)) {
|
||||
expect(record[key]).toEqual(expected);
|
||||
}
|
||||
return record;
|
||||
}
|
||||
|
||||
function gatewayCall(index: number) {
|
||||
const call = callGatewayFromCli.mock.calls[index];
|
||||
if (!call) {
|
||||
throw new Error(`Expected gateway call ${index + 1}`);
|
||||
}
|
||||
return call;
|
||||
}
|
||||
|
||||
function expectGatewayCall(index: number, method: string, params: unknown) {
|
||||
const call = gatewayCall(index);
|
||||
expect(call[0]).toBe(method);
|
||||
expect(requireRecord(call[1], "gateway call options").timeout).toBe("60000");
|
||||
expect(call[2]).toEqual(params);
|
||||
}
|
||||
|
||||
function writtenJson(): Record<string, unknown> {
|
||||
const value = vi.mocked(defaultRuntime.writeJson).mock.calls[0]?.[0];
|
||||
return requireRecord(value, "written json");
|
||||
}
|
||||
|
||||
function effectivePolicy(output: Record<string, unknown> = writtenJson()) {
|
||||
return requireRecord(output.effectivePolicy, "effective policy");
|
||||
}
|
||||
|
||||
function scopes(output: Record<string, unknown> = writtenJson()) {
|
||||
return requireArray(effectivePolicy(output).scopes, "effective policy scopes");
|
||||
}
|
||||
|
||||
function scopeByLabel(label: string, output: Record<string, unknown> = writtenJson()) {
|
||||
const scope = scopes(output).find(
|
||||
(entry) => requireRecord(entry, "policy scope").scopeLabel === label,
|
||||
);
|
||||
if (!scope) {
|
||||
throw new Error(`Expected policy scope ${label}`);
|
||||
}
|
||||
return requireRecord(scope, `policy scope ${label}`);
|
||||
}
|
||||
|
||||
function resetLocalSnapshot() {
|
||||
localSnapshot.file = { version: 1, agents: {} };
|
||||
}
|
||||
@@ -202,15 +138,35 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--gateway"]);
|
||||
|
||||
expectGatewayCall(0, "exec.approvals.get", {});
|
||||
expectGatewayCall(1, "config.get", {});
|
||||
expect(callGatewayFromCli).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
"exec.approvals.get",
|
||||
expect.objectContaining({ timeout: "60000" }),
|
||||
{},
|
||||
);
|
||||
expect(callGatewayFromCli).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
"config.get",
|
||||
expect.objectContaining({ timeout: "60000" }),
|
||||
{},
|
||||
);
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
callGatewayFromCli.mockClear();
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--node", "macbook"]);
|
||||
|
||||
expectGatewayCall(0, "exec.approvals.node.get", { nodeId: "node-1" });
|
||||
expectGatewayCall(1, "config.get", {});
|
||||
expect(callGatewayFromCli).toHaveBeenCalledWith(
|
||||
"exec.approvals.node.get",
|
||||
expect.objectContaining({ timeout: "60000" }),
|
||||
{
|
||||
nodeId: "node-1",
|
||||
},
|
||||
);
|
||||
expect(callGatewayFromCli).toHaveBeenCalledWith(
|
||||
"config.get",
|
||||
expect.objectContaining({ timeout: "60000" }),
|
||||
{},
|
||||
);
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
});
|
||||
|
||||
@@ -231,22 +187,29 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
const policy = effectivePolicy();
|
||||
expect(policy.note).toBe(
|
||||
"Effective exec policy is the host approvals file intersected with requested tools.exec policy.",
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: {
|
||||
note: "Effective exec policy is the host approvals file intersected with requested tools.exec policy.",
|
||||
scopes: [
|
||||
expect.objectContaining({
|
||||
scopeLabel: "tools.exec",
|
||||
security: expect.objectContaining({
|
||||
requested: "full",
|
||||
host: "allowlist",
|
||||
effective: "allowlist",
|
||||
}),
|
||||
ask: expect.objectContaining({
|
||||
requested: "off",
|
||||
host: "always",
|
||||
effective: "always",
|
||||
}),
|
||||
}),
|
||||
],
|
||||
},
|
||||
}),
|
||||
0,
|
||||
);
|
||||
const scope = scopeByLabel("tools.exec");
|
||||
expectFields(requireRecord(scope.security, "tools.exec security"), "tools.exec security", {
|
||||
requested: "full",
|
||||
host: "allowlist",
|
||||
effective: "allowlist",
|
||||
});
|
||||
expectFields(requireRecord(scope.ask, "tools.exec ask"), "tools.exec ask", {
|
||||
requested: "off",
|
||||
host: "always",
|
||||
effective: "always",
|
||||
});
|
||||
});
|
||||
|
||||
it("reports wildcard host policy sources in effective policy output", async () => {
|
||||
@@ -279,16 +242,26 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
const scope = scopeByLabel("agent:runner");
|
||||
expect(requireRecord(scope.security, "agent security").hostSource).toBe(
|
||||
"/tmp/local-exec-approvals.json agents.*.security",
|
||||
);
|
||||
expect(requireRecord(scope.ask, "agent ask").hostSource).toBe(
|
||||
"/tmp/local-exec-approvals.json agents.*.ask",
|
||||
);
|
||||
expect(requireRecord(scope.askFallback, "agent askFallback").source).toBe(
|
||||
"/tmp/local-exec-approvals.json agents.*.askFallback",
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: expect.objectContaining({
|
||||
scopes: expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
scopeLabel: "agent:runner",
|
||||
security: expect.objectContaining({
|
||||
hostSource: "/tmp/local-exec-approvals.json agents.*.security",
|
||||
}),
|
||||
ask: expect.objectContaining({
|
||||
hostSource: "/tmp/local-exec-approvals.json agents.*.ask",
|
||||
}),
|
||||
askFallback: expect.objectContaining({
|
||||
source: "/tmp/local-exec-approvals.json agents.*.askFallback",
|
||||
}),
|
||||
}),
|
||||
]),
|
||||
}),
|
||||
}),
|
||||
0,
|
||||
);
|
||||
});
|
||||
|
||||
@@ -325,29 +298,32 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--node", "macbook", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
const policy = effectivePolicy();
|
||||
expect(policy.note).toBe(
|
||||
"Effective exec policy is the node host approvals file intersected with gateway tools.exec policy.",
|
||||
);
|
||||
const scope = scopeByLabel("tools.exec");
|
||||
expectFields(requireRecord(scope.security, "tools.exec security"), "tools.exec security", {
|
||||
requested: "full",
|
||||
host: "allowlist",
|
||||
effective: "allowlist",
|
||||
});
|
||||
expectFields(requireRecord(scope.ask, "tools.exec ask"), "tools.exec ask", {
|
||||
requested: "off",
|
||||
host: "always",
|
||||
effective: "always",
|
||||
});
|
||||
expectFields(
|
||||
requireRecord(scope.askFallback, "tools.exec askFallback"),
|
||||
"tools.exec askFallback",
|
||||
{
|
||||
effective: "deny",
|
||||
source: "/tmp/node-exec-approvals.json defaults.askFallback",
|
||||
},
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: {
|
||||
note: "Effective exec policy is the node host approvals file intersected with gateway tools.exec policy.",
|
||||
scopes: [
|
||||
expect.objectContaining({
|
||||
scopeLabel: "tools.exec",
|
||||
security: expect.objectContaining({
|
||||
requested: "full",
|
||||
host: "allowlist",
|
||||
effective: "allowlist",
|
||||
}),
|
||||
ask: expect.objectContaining({
|
||||
requested: "off",
|
||||
host: "always",
|
||||
effective: "always",
|
||||
}),
|
||||
askFallback: expect.objectContaining({
|
||||
effective: "deny",
|
||||
source: "/tmp/node-exec-approvals.json defaults.askFallback",
|
||||
}),
|
||||
}),
|
||||
],
|
||||
},
|
||||
}),
|
||||
0,
|
||||
);
|
||||
});
|
||||
|
||||
@@ -371,11 +347,15 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--gateway", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
expect(effectivePolicy()).toEqual({
|
||||
note: "Config unavailable.",
|
||||
scopes: [],
|
||||
});
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: {
|
||||
note: "Config unavailable.",
|
||||
scopes: [],
|
||||
},
|
||||
}),
|
||||
0,
|
||||
);
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
});
|
||||
|
||||
@@ -399,11 +379,15 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--gateway", "--timeout", "10000", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
expect(effectivePolicy()).toEqual({
|
||||
note: "Config fetch timed out. Re-run with a higher --timeout to inspect Effective Policy.",
|
||||
scopes: [],
|
||||
});
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: {
|
||||
note: "Config fetch timed out. Re-run with a higher --timeout to inspect Effective Policy.",
|
||||
scopes: [],
|
||||
},
|
||||
}),
|
||||
0,
|
||||
);
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
});
|
||||
|
||||
@@ -427,11 +411,15 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--node", "macbook", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
expect(effectivePolicy()).toEqual({
|
||||
note: "Gateway config unavailable. Node output above shows host approvals state only, and final runtime policy still intersects with gateway tools.exec.",
|
||||
scopes: [],
|
||||
});
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: {
|
||||
note: "Gateway config unavailable. Node output above shows host approvals state only, and final runtime policy still intersects with gateway tools.exec.",
|
||||
scopes: [],
|
||||
},
|
||||
}),
|
||||
0,
|
||||
);
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
});
|
||||
|
||||
@@ -440,11 +428,15 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "get", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
expect(effectivePolicy()).toEqual({
|
||||
note: "Config unavailable.",
|
||||
scopes: [],
|
||||
});
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
effectivePolicy: {
|
||||
note: "Config unavailable.",
|
||||
scopes: [],
|
||||
},
|
||||
}),
|
||||
0,
|
||||
);
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
});
|
||||
|
||||
@@ -473,43 +465,50 @@ describe("exec approvals CLI", () => {
|
||||
await runApprovalsCommand(["approvals", "get", "--json"]);
|
||||
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledTimes(1);
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(writtenJson(), 0);
|
||||
expect(defaultRuntime.writeJson).toHaveBeenCalledWith(expect.anything(), 0);
|
||||
|
||||
const toolsScope = scopeByLabel("tools.exec");
|
||||
expectFields(requireRecord(toolsScope.security, "tools.exec security"), "tools.exec security", {
|
||||
requested: "full",
|
||||
requestedSource: "tools.exec.security",
|
||||
effective: "full",
|
||||
});
|
||||
expectFields(requireRecord(toolsScope.ask, "tools.exec ask"), "tools.exec ask", {
|
||||
requested: "off",
|
||||
requestedSource: "tools.exec.ask",
|
||||
effective: "off",
|
||||
});
|
||||
expectFields(
|
||||
requireRecord(toolsScope.askFallback, "tools.exec askFallback"),
|
||||
"tools.exec askFallback",
|
||||
{
|
||||
effective: "full",
|
||||
source: "OpenClaw default (full)",
|
||||
},
|
||||
const output = vi.mocked(defaultRuntime.writeJson).mock.calls[0]?.[0] as {
|
||||
effectivePolicy: { scopes: unknown[] };
|
||||
};
|
||||
|
||||
expect(output.effectivePolicy.scopes).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({
|
||||
scopeLabel: "tools.exec",
|
||||
security: expect.objectContaining({
|
||||
requested: "full",
|
||||
requestedSource: "tools.exec.security",
|
||||
effective: "full",
|
||||
}),
|
||||
ask: expect.objectContaining({
|
||||
requested: "off",
|
||||
requestedSource: "tools.exec.ask",
|
||||
effective: "off",
|
||||
}),
|
||||
askFallback: expect.objectContaining({
|
||||
effective: "full",
|
||||
source: "OpenClaw default (full)",
|
||||
}),
|
||||
}),
|
||||
expect.objectContaining({
|
||||
scopeLabel: "agent:runner",
|
||||
security: expect.objectContaining({
|
||||
requested: "full",
|
||||
requestedSource: "tools.exec.security",
|
||||
effective: "allowlist",
|
||||
}),
|
||||
ask: expect.objectContaining({
|
||||
requested: "off",
|
||||
requestedSource: "tools.exec.ask",
|
||||
effective: "always",
|
||||
}),
|
||||
askFallback: expect.objectContaining({
|
||||
effective: "allowlist",
|
||||
source: "OpenClaw default (full)",
|
||||
}),
|
||||
}),
|
||||
]),
|
||||
);
|
||||
|
||||
const agentScope = scopeByLabel("agent:runner");
|
||||
expectFields(requireRecord(agentScope.security, "agent security"), "agent security", {
|
||||
requested: "full",
|
||||
requestedSource: "tools.exec.security",
|
||||
effective: "allowlist",
|
||||
});
|
||||
expectFields(requireRecord(agentScope.ask, "agent ask"), "agent ask", {
|
||||
requested: "off",
|
||||
requestedSource: "tools.exec.ask",
|
||||
effective: "always",
|
||||
});
|
||||
expectFields(requireRecord(agentScope.askFallback, "agent askFallback"), "agent askFallback", {
|
||||
effective: "allowlist",
|
||||
source: "OpenClaw default (full)",
|
||||
});
|
||||
});
|
||||
|
||||
it("defaults allowlist add to wildcard agent", async () => {
|
||||
@@ -518,14 +517,18 @@ describe("exec approvals CLI", () => {
|
||||
|
||||
await runApprovalsCommand(["approvals", "allowlist", "add", "/usr/bin/uname"]);
|
||||
|
||||
expect(callGatewayFromCli.mock.calls.some((call) => call[0] === "exec.approvals.set")).toBe(
|
||||
false,
|
||||
expect(callGatewayFromCli).not.toHaveBeenCalledWith(
|
||||
"exec.approvals.set",
|
||||
expect.anything(),
|
||||
{},
|
||||
);
|
||||
expect(saveExecApprovals).toHaveBeenCalledWith(
|
||||
requireRecord(saveExecApprovals.mock.calls[0]?.[0], "saved approvals"),
|
||||
expect.objectContaining({
|
||||
agents: expect.objectContaining({
|
||||
"*": expect.anything(),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
const saved = requireRecord(saveExecApprovals.mock.calls[0]?.[0], "saved approvals");
|
||||
expect(requireRecord(saved.agents, "saved agents")["*"]).toBeDefined();
|
||||
});
|
||||
|
||||
it("removes wildcard allowlist entry and prunes empty agent", async () => {
|
||||
@@ -544,12 +547,11 @@ describe("exec approvals CLI", () => {
|
||||
await runApprovalsCommand(["approvals", "allowlist", "remove", "/usr/bin/uname"]);
|
||||
|
||||
expect(saveExecApprovals).toHaveBeenCalledWith(
|
||||
requireRecord(saveExecApprovals.mock.calls[0]?.[0], "saved approvals"),
|
||||
expect.objectContaining({
|
||||
version: 1,
|
||||
agents: undefined,
|
||||
}),
|
||||
);
|
||||
expectFields(saveExecApprovals.mock.calls[0]?.[0], "saved approvals", {
|
||||
version: 1,
|
||||
agents: undefined,
|
||||
});
|
||||
expect(runtimeErrors).toHaveLength(0);
|
||||
});
|
||||
});
|
||||
|
||||
@@ -19,8 +19,7 @@ const forceFreePortAndWait = vi.fn(async (_port: number, _opts: unknown) => ({
|
||||
}));
|
||||
const waitForPortBindable = vi.fn(async (_port: number, _opts?: unknown) => 0);
|
||||
const ensureDevGatewayConfig = vi.fn(async (_opts?: unknown) => {});
|
||||
type GatewayLoopStart = (params?: { startupStartedAt?: number }) => Promise<unknown>;
|
||||
const runGatewayLoop = vi.fn(async ({ start }: { start: GatewayLoopStart }) => {
|
||||
const runGatewayLoop = vi.fn(async ({ start }: { start: () => Promise<unknown> }) => {
|
||||
await start();
|
||||
});
|
||||
const gatewayLogMessages = vi.hoisted(() => [] as string[]);
|
||||
@@ -158,7 +157,7 @@ vi.mock("./dev.js", () => ({
|
||||
}));
|
||||
|
||||
vi.mock("./run-loop.js", () => ({
|
||||
runGatewayLoop: (params: { start: GatewayLoopStart }) => runGatewayLoop(params),
|
||||
runGatewayLoop: (params: { start: () => Promise<unknown> }) => runGatewayLoop(params),
|
||||
}));
|
||||
|
||||
describe("gateway run option collisions", () => {
|
||||
@@ -306,41 +305,6 @@ describe("gateway run option collisions", () => {
|
||||
);
|
||||
});
|
||||
|
||||
it("uses the startup snapshot only for the first in-process gateway start", async () => {
|
||||
runGatewayLoop.mockImplementationOnce(async ({ start }: { start: GatewayLoopStart }) => {
|
||||
await start({ startupStartedAt: 1000 });
|
||||
await start({ startupStartedAt: 2000 });
|
||||
});
|
||||
|
||||
await runGatewayCli(["gateway", "run", "--allow-unconfigured"]);
|
||||
|
||||
expect(startGatewayServer).toHaveBeenCalledTimes(2);
|
||||
expect(startGatewayServer).toHaveBeenNthCalledWith(
|
||||
1,
|
||||
18789,
|
||||
expect.objectContaining({
|
||||
startupStartedAt: 1000,
|
||||
startupConfigSnapshotRead: {
|
||||
snapshot: configState.snapshot,
|
||||
},
|
||||
}),
|
||||
);
|
||||
expect(startGatewayServer).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
18789,
|
||||
expect.not.objectContaining({
|
||||
startupConfigSnapshotRead: expect.anything(),
|
||||
}),
|
||||
);
|
||||
expect(startGatewayServer).toHaveBeenNthCalledWith(
|
||||
2,
|
||||
18789,
|
||||
expect.objectContaining({
|
||||
startupStartedAt: 2000,
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
it("logs when first startup will build missing Control UI assets", async () => {
|
||||
controlUiState.root = null;
|
||||
|
||||
|
||||
@@ -787,25 +787,19 @@ async function runGatewayCommand(opts: GatewayRunOpts) {
|
||||
gatewayLog.info("starting...");
|
||||
startupTrace.mark("cli.gateway-loop");
|
||||
const healthHost = await resolveGatewayBindHost(bind, cfg.gateway?.customBindHost);
|
||||
let startupConfigSnapshotReadForNextStart = startupConfigSnapshotRead;
|
||||
const startLoop = async () =>
|
||||
await runGatewayLoop({
|
||||
runtime: defaultRuntime,
|
||||
lockPort: port,
|
||||
healthHost,
|
||||
start: async ({ startupStartedAt } = {}) => {
|
||||
const startupConfigSnapshotReadForThisStart = startupConfigSnapshotReadForNextStart;
|
||||
startupConfigSnapshotReadForNextStart = undefined;
|
||||
return await startGatewayServer(port, {
|
||||
start: async ({ startupStartedAt } = {}) =>
|
||||
await startGatewayServer(port, {
|
||||
bind,
|
||||
auth: authOverride,
|
||||
tailscale: tailscaleOverride,
|
||||
startupStartedAt,
|
||||
...(startupConfigSnapshotReadForThisStart
|
||||
? { startupConfigSnapshotRead: startupConfigSnapshotReadForThisStart }
|
||||
: {}),
|
||||
});
|
||||
},
|
||||
...(startupConfigSnapshotRead ? { startupConfigSnapshotRead } : {}),
|
||||
}),
|
||||
});
|
||||
|
||||
const { detectRespawnSupervisor } = await import("../../infra/supervisor-markers.js");
|
||||
|
||||
@@ -112,11 +112,6 @@ export function resolveMissingPluginCommandMessage(
|
||||
config?: OpenClawConfig;
|
||||
registry?: PluginManifestCommandAliasRegistry;
|
||||
}) => PluginManifestToolOwnerRecord | undefined;
|
||||
resolveCliCommandSurfaceOwner?: (params: {
|
||||
command: string | undefined;
|
||||
config?: OpenClawConfig;
|
||||
registry?: PluginManifestCommandAliasRegistry;
|
||||
}) => string | undefined;
|
||||
},
|
||||
): string | null {
|
||||
const normalizedPluginId = normalizeLowercaseStringOrEmpty(pluginId);
|
||||
@@ -235,33 +230,6 @@ export function resolveMissingPluginCommandMessage(
|
||||
if (parentPluginId && allow.includes(parentPluginId)) {
|
||||
return null;
|
||||
}
|
||||
const cliCommandSurfaceOwner = options?.resolveCliCommandSurfaceOwner
|
||||
? options.resolveCliCommandSurfaceOwner({
|
||||
command: normalizedPluginId,
|
||||
config,
|
||||
...(options?.registry ? { registry: options.registry } : {}),
|
||||
})
|
||||
: options?.registry
|
||||
? resolveManifestCommandAliasOwnerInRegistry({
|
||||
command: normalizedPluginId,
|
||||
registry: options.registry,
|
||||
})?.pluginId
|
||||
: undefined;
|
||||
const normalizedCliCommandSurfaceOwner =
|
||||
normalizeOptionalLowercaseString(cliCommandSurfaceOwner);
|
||||
if (!normalizedCliCommandSurfaceOwner) {
|
||||
return null;
|
||||
}
|
||||
if (allow.includes(normalizedCliCommandSurfaceOwner)) {
|
||||
return null;
|
||||
}
|
||||
if (normalizedCliCommandSurfaceOwner !== normalizedPluginId) {
|
||||
return (
|
||||
`"${normalizedPluginId}" is not a plugin; it is a command provided by the ` +
|
||||
`"${normalizedCliCommandSurfaceOwner}" plugin. Add "${normalizedCliCommandSurfaceOwner}" to ` +
|
||||
`\`plugins.allow\` instead of "${normalizedPluginId}".`
|
||||
);
|
||||
}
|
||||
return (
|
||||
`The \`openclaw ${normalizedPluginId}\` command is unavailable because ` +
|
||||
`\`plugins.allow\` excludes "${normalizedPluginId}". Add "${normalizedPluginId}" to ` +
|
||||
|
||||
@@ -25,7 +25,6 @@ const registerPluginCliCommandsFromValidatedConfigMock = vi.hoisted(() => vi.fn(
|
||||
const resolvePluginCliRootOwnerIdsMock = vi.hoisted(() => vi.fn());
|
||||
const resolveManifestCommandAliasOwnerMock = vi.hoisted(() => vi.fn());
|
||||
const resolveManifestToolOwnerMock = vi.hoisted(() => vi.fn());
|
||||
const resolveManifestCliCommandSurfaceOwnerMock = vi.hoisted(() => vi.fn());
|
||||
const restoreTerminalStateMock = vi.hoisted(() => vi.fn());
|
||||
const hasEnvHttpProxyAgentConfiguredMock = vi.hoisted(() => vi.fn(() => false));
|
||||
const ensureGlobalUndiciEnvProxyDispatcherMock = vi.hoisted(() => vi.fn());
|
||||
@@ -180,7 +179,6 @@ vi.mock("../plugins/cli-registry-loader.js", () => ({
|
||||
}));
|
||||
|
||||
vi.mock("../plugins/manifest-command-aliases.runtime.js", () => ({
|
||||
resolveManifestCliCommandSurfaceOwner: resolveManifestCliCommandSurfaceOwnerMock,
|
||||
resolveManifestCommandAliasOwner: resolveManifestCommandAliasOwnerMock,
|
||||
resolveManifestToolOwner: resolveManifestToolOwnerMock,
|
||||
}));
|
||||
@@ -254,7 +252,6 @@ describe("runCli exit behavior", () => {
|
||||
);
|
||||
resolveManifestCommandAliasOwnerMock.mockReturnValue(undefined);
|
||||
resolveManifestToolOwnerMock.mockReturnValue(undefined);
|
||||
resolveManifestCliCommandSurfaceOwnerMock.mockReturnValue(undefined);
|
||||
delete process.env.OPENCLAW_DISABLE_CLI_STARTUP_HELP_FAST_PATH;
|
||||
delete process.env.OPENCLAW_HIDE_BANNER;
|
||||
});
|
||||
@@ -491,53 +488,6 @@ describe("runCli exit behavior", () => {
|
||||
expect(registerPluginCliCommandsFromValidatedConfigMock).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it("does not suggest plugins.allow for unknown command roots before proxy startup", async () => {
|
||||
loadConfigMock.mockReturnValueOnce({
|
||||
plugins: {
|
||||
allow: ["browser"],
|
||||
},
|
||||
});
|
||||
|
||||
let error: unknown;
|
||||
try {
|
||||
await runCli(["node", "openclaw", "totally-unknown"]);
|
||||
} catch (caught) {
|
||||
error = caught;
|
||||
}
|
||||
expect(error).toBeInstanceOf(Error);
|
||||
expect((error as Error).message).toContain(
|
||||
'No built-in command or plugin CLI metadata owns "totally-unknown"',
|
||||
);
|
||||
expect((error as Error).message).not.toContain("plugins.allow");
|
||||
expect(startProxyMock).not.toHaveBeenCalled();
|
||||
expect(tryRouteCliMock).not.toHaveBeenCalled();
|
||||
expect(registerPluginCliCommandsFromValidatedConfigMock).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it("preserves plugins.allow diagnostics for roots owned only by CLI metadata", async () => {
|
||||
loadConfigMock.mockReturnValueOnce({
|
||||
plugins: {
|
||||
allow: ["browser"],
|
||||
},
|
||||
});
|
||||
resolvePluginCliRootOwnerIdsMock.mockImplementation(
|
||||
({
|
||||
cfg,
|
||||
primaryCommand,
|
||||
}: {
|
||||
cfg?: { plugins?: { allow?: string[] } };
|
||||
primaryCommand?: string;
|
||||
}) => (primaryCommand === "qa" && cfg?.plugins?.allow?.length === 0 ? ["qa-lab"] : []),
|
||||
);
|
||||
|
||||
await expect(runCli(["node", "openclaw", "qa"])).rejects.toThrow(
|
||||
'Add "qa-lab" to `plugins.allow` instead of "qa"',
|
||||
);
|
||||
expect(startProxyMock).not.toHaveBeenCalled();
|
||||
expect(tryRouteCliMock).not.toHaveBeenCalled();
|
||||
expect(registerPluginCliCommandsFromValidatedConfigMock).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it("reports plugin tool command mistakes before proxy startup", async () => {
|
||||
resolveManifestToolOwnerMock.mockReturnValueOnce({
|
||||
toolName: "lcm_recent",
|
||||
|
||||
@@ -40,16 +40,6 @@ const losslessClawToolRegistry: PluginManifestCommandAliasRegistry = {
|
||||
],
|
||||
};
|
||||
|
||||
const browserCommandAliasRegistry: PluginManifestCommandAliasRegistry = {
|
||||
plugins: [
|
||||
{
|
||||
id: "browser",
|
||||
enabledByDefault: true,
|
||||
commandAliases: [{ name: "browser" }],
|
||||
},
|
||||
],
|
||||
};
|
||||
|
||||
describe("isGatewayRunFastPathArgv", () => {
|
||||
it("matches only plain gateway foreground starts without root options or help", () => {
|
||||
expect(isGatewayRunFastPathArgv(["node", "openclaw", "gateway"])).toBe(true);
|
||||
@@ -201,15 +191,11 @@ describe("shouldUseBrowserHelpFastPath", () => {
|
||||
describe("resolveMissingPluginCommandMessage", () => {
|
||||
it("explains plugins.allow misses for a bundled plugin command", () => {
|
||||
expect(
|
||||
resolveMissingPluginCommandMessage(
|
||||
"browser",
|
||||
{
|
||||
plugins: {
|
||||
allow: ["quietchat"],
|
||||
},
|
||||
resolveMissingPluginCommandMessage("browser", {
|
||||
plugins: {
|
||||
allow: ["quietchat"],
|
||||
},
|
||||
{ registry: browserCommandAliasRegistry },
|
||||
),
|
||||
}),
|
||||
).toContain('`plugins.allow` excludes "browser"');
|
||||
});
|
||||
|
||||
@@ -416,7 +402,7 @@ describe("resolveMissingPluginCommandMessage", () => {
|
||||
expect(message).toContain('"lossless-claw"');
|
||||
});
|
||||
|
||||
it("returns null for unknown names excluded by plugins.allow", () => {
|
||||
it("preserves the plugins.allow suggestion when the unknown name is not a plugin tool", () => {
|
||||
const message = resolveMissingPluginCommandMessage(
|
||||
"totally-unknown",
|
||||
{
|
||||
@@ -426,30 +412,14 @@ describe("resolveMissingPluginCommandMessage", () => {
|
||||
},
|
||||
{ registry: losslessClawToolRegistry },
|
||||
);
|
||||
expect(message).toBeNull();
|
||||
});
|
||||
|
||||
it("points metadata-only CLI roots in plugins.allow at their parent plugin", () => {
|
||||
const message = resolveMissingPluginCommandMessage(
|
||||
"qa",
|
||||
{
|
||||
plugins: {
|
||||
allow: ["browser"],
|
||||
},
|
||||
},
|
||||
{
|
||||
resolveCliCommandSurfaceOwner: () => "qa-lab",
|
||||
},
|
||||
);
|
||||
expect(message).toContain('"qa" is not a plugin');
|
||||
expect(message).toContain('"qa-lab"');
|
||||
expect(message).toContain('Add "qa-lab" to `plugins.allow` instead of "qa"');
|
||||
expect(message).not.toBeNull();
|
||||
expect(message).toContain('`plugins.allow` excludes "totally-unknown"');
|
||||
});
|
||||
|
||||
it("does not attribute a tool to an owning plugin excluded by plugins.allow", () => {
|
||||
// The owning plugin is denied via plugins.allow, so the manifest-declared
|
||||
// tool is not available through the owning plugin. Tool names are not CLI
|
||||
// command surfaces, so do not suggest adding the tool name to plugins.allow.
|
||||
// tool is not available through the owning plugin. Fall through to the
|
||||
// standard plugins.allow message instead of falsely attributing it.
|
||||
const message = resolveMissingPluginCommandMessage(
|
||||
"lcm_recent",
|
||||
{
|
||||
@@ -459,7 +429,9 @@ describe("resolveMissingPluginCommandMessage", () => {
|
||||
},
|
||||
{ registry: losslessClawToolRegistry },
|
||||
);
|
||||
expect(message).toBeNull();
|
||||
expect(message).not.toBeNull();
|
||||
expect(message).not.toContain("agent tool available");
|
||||
expect(message).toContain('`plugins.allow` excludes "lcm_recent"');
|
||||
});
|
||||
|
||||
it("does not attribute a tool to an owning plugin disabled via plugins.entries", () => {
|
||||
|
||||
@@ -319,47 +319,6 @@ async function isPluginCliRoot(params: {
|
||||
}
|
||||
}
|
||||
|
||||
function createAllowlistAgnosticCliLookupConfig(config: OpenClawConfig): OpenClawConfig {
|
||||
if (!Array.isArray(config.plugins?.allow) || config.plugins.allow.length === 0) {
|
||||
return config;
|
||||
}
|
||||
return {
|
||||
...config,
|
||||
plugins: {
|
||||
...config.plugins,
|
||||
allow: [],
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
async function resolveCliCommandSurfaceOwner(params: {
|
||||
primary: string;
|
||||
config: OpenClawConfig;
|
||||
}): Promise<string | undefined> {
|
||||
const { resolveManifestCliCommandSurfaceOwner } =
|
||||
await import("../plugins/manifest-command-aliases.runtime.js");
|
||||
const manifestOwner = resolveManifestCliCommandSurfaceOwner({
|
||||
command: params.primary,
|
||||
config: params.config,
|
||||
env: process.env,
|
||||
});
|
||||
if (manifestOwner) {
|
||||
return manifestOwner;
|
||||
}
|
||||
try {
|
||||
const { resolvePluginCliRootOwnerIds } = await import("../plugins/cli-registry-loader.js");
|
||||
return (
|
||||
await resolvePluginCliRootOwnerIds({
|
||||
cfg: createAllowlistAgnosticCliLookupConfig(params.config),
|
||||
env: process.env,
|
||||
primaryCommand: params.primary,
|
||||
})
|
||||
)?.[0];
|
||||
} catch {
|
||||
return undefined;
|
||||
}
|
||||
}
|
||||
|
||||
async function resolveUnownedCliPrimary(params: {
|
||||
argv: string[];
|
||||
config: OpenClawConfig;
|
||||
@@ -388,12 +347,10 @@ async function resolveUnownedCliPrimaryMessage(params: {
|
||||
}): Promise<string> {
|
||||
const { resolveManifestCommandAliasOwner, resolveManifestToolOwner } =
|
||||
await import("../plugins/manifest-command-aliases.runtime.js");
|
||||
const cliCommandSurfaceOwner = await resolveCliCommandSurfaceOwner(params);
|
||||
return (
|
||||
resolveMissingPluginCommandMessageFromPolicy(params.primary, params.config, {
|
||||
resolveCommandAliasOwner: resolveManifestCommandAliasOwner,
|
||||
resolveToolOwner: resolveManifestToolOwner,
|
||||
resolveCliCommandSurfaceOwner: () => cliCommandSurfaceOwner,
|
||||
}) ??
|
||||
`Unknown command: openclaw ${params.primary}. No built-in command or plugin CLI metadata owns "${params.primary}".`
|
||||
);
|
||||
@@ -739,17 +696,12 @@ export async function runCli(argv: string[] = process.argv) {
|
||||
) {
|
||||
const { resolveManifestCommandAliasOwner, resolveManifestToolOwner } =
|
||||
await import("../plugins/manifest-command-aliases.runtime.js");
|
||||
const cliCommandSurfaceOwner = await resolveCliCommandSurfaceOwner({
|
||||
primary,
|
||||
config,
|
||||
});
|
||||
const missingPluginCommandMessage = resolveMissingPluginCommandMessageFromPolicy(
|
||||
primary,
|
||||
config,
|
||||
{
|
||||
resolveCommandAliasOwner: resolveManifestCommandAliasOwner,
|
||||
resolveToolOwner: resolveManifestToolOwner,
|
||||
resolveCliCommandSurfaceOwner: () => cliCommandSurfaceOwner,
|
||||
},
|
||||
);
|
||||
if (missingPluginCommandMessage) {
|
||||
|
||||
@@ -194,76 +194,6 @@ function createGatewayCommand(entrypoint: string) {
|
||||
};
|
||||
}
|
||||
|
||||
function requireRecord(value: unknown, label: string): Record<string, unknown> {
|
||||
expect(value, label).toBeTypeOf("object");
|
||||
expect(value, label).not.toBeNull();
|
||||
return value as Record<string, unknown>;
|
||||
}
|
||||
|
||||
function callArg(mock: { mock: { calls: Array<Array<unknown>> } }, index: number, label: string) {
|
||||
const call = mock.mock.calls.at(index);
|
||||
expect(call, label).toBeDefined();
|
||||
return call?.[0];
|
||||
}
|
||||
|
||||
function expectCallField(
|
||||
mock: { mock: { calls: Array<Array<unknown>> } },
|
||||
field: string,
|
||||
expected: unknown,
|
||||
) {
|
||||
const options = requireRecord(callArg(mock, 0, `first ${field} call`), field);
|
||||
expect(options[field]).toEqual(expected);
|
||||
return options;
|
||||
}
|
||||
|
||||
function expectGatewayAuthToken(value: unknown, expected: string) {
|
||||
const root = requireRecord(value, "config root");
|
||||
const gateway = requireRecord(root.gateway, "config.gateway");
|
||||
const auth = requireRecord(gateway.auth, "config.gateway.auth");
|
||||
expect(auth.token).toBe(expected);
|
||||
}
|
||||
|
||||
function readGatewayAuthToken(value: unknown) {
|
||||
if (!value || typeof value !== "object") {
|
||||
return undefined;
|
||||
}
|
||||
const root = value as Record<string, unknown>;
|
||||
const gateway = root.gateway;
|
||||
if (!gateway || typeof gateway !== "object") {
|
||||
return undefined;
|
||||
}
|
||||
const auth = (gateway as Record<string, unknown>).auth;
|
||||
if (!auth || typeof auth !== "object") {
|
||||
return undefined;
|
||||
}
|
||||
return (auth as Record<string, unknown>).token;
|
||||
}
|
||||
|
||||
function expectCallConfigGatewayAuthToken(
|
||||
mock: { mock: { calls: Array<Array<unknown>> } },
|
||||
expected: string,
|
||||
) {
|
||||
const matched = mock.mock.calls.some(([value]) => {
|
||||
const options = value && typeof value === "object" ? (value as Record<string, unknown>) : {};
|
||||
return readGatewayAuthToken(options.config) === expected;
|
||||
});
|
||||
expect(matched).toBe(true);
|
||||
}
|
||||
|
||||
function expectNoteContaining(messagePart: string, title: string) {
|
||||
const messages = mocks.note.mock.calls
|
||||
.filter(([, callTitle]) => callTitle === title)
|
||||
.map(([message]) => String(message));
|
||||
expect(messages.some((message) => message.includes(messagePart))).toBe(true);
|
||||
}
|
||||
|
||||
function expectNoNoteContaining(messagePart: string, title: string) {
|
||||
const messages = mocks.note.mock.calls
|
||||
.filter(([, callTitle]) => callTitle === title)
|
||||
.map(([message]) => String(message));
|
||||
expect(messages.some((message) => message.includes(messagePart))).toBe(false);
|
||||
}
|
||||
|
||||
function setupGatewayEntrypointRepairScenario(params: {
|
||||
currentEntrypoint: string;
|
||||
installEntrypoint: string;
|
||||
@@ -358,8 +288,22 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair(cfg);
|
||||
|
||||
expectCallField(mocks.auditGatewayServiceConfig, "expectedGatewayToken", "config-token");
|
||||
expectCallConfigGatewayAuthToken(mocks.buildGatewayInstallPlan, "config-token");
|
||||
expect(mocks.auditGatewayServiceConfig).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
expectedGatewayToken: "config-token",
|
||||
}),
|
||||
);
|
||||
expect(mocks.buildGatewayInstallPlan).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
config: expect.objectContaining({
|
||||
gateway: expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
token: "config-token",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(mocks.replaceConfigFile).not.toHaveBeenCalled();
|
||||
expect(mocks.stage).not.toHaveBeenCalled();
|
||||
expect(mocks.install).toHaveBeenCalledTimes(1);
|
||||
@@ -399,12 +343,14 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
const runtimeNotes = mocks.note.mock.calls.filter(([, title]) => title === "Gateway runtime");
|
||||
const runtimeMessages = runtimeNotes.map(([message]) => message);
|
||||
expect(runtimeMessages).not.toContain("duplicate doctor runtime warning");
|
||||
expect(runtimeMessages.some((message) => String(message).includes("not found"))).toBe(false);
|
||||
expect(
|
||||
runtimeMessages.some((message) =>
|
||||
String(message).includes("Using /home/orin/.nvm/versions/node/v22.22.2/bin/node"),
|
||||
),
|
||||
).toBe(true);
|
||||
expect(runtimeMessages).not.toEqual(
|
||||
expect.arrayContaining([expect.stringContaining("not found")]),
|
||||
);
|
||||
expect(runtimeMessages).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.stringContaining("Using /home/orin/.nvm/versions/node/v22.22.2/bin/node"),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
it("passes planned managed env keys into service audit for legacy inline secret detection", async () => {
|
||||
@@ -436,10 +382,10 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair({ gateway: {} });
|
||||
|
||||
expectCallField(
|
||||
mocks.auditGatewayServiceConfig,
|
||||
"expectedManagedServiceEnvKeys",
|
||||
new Set(["TAVILY_API_KEY"]),
|
||||
expect(mocks.auditGatewayServiceConfig).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
expectedManagedServiceEnvKeys: new Set(["TAVILY_API_KEY"]),
|
||||
}),
|
||||
);
|
||||
expect(mocks.install).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
@@ -470,12 +416,16 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair({ gateway: { port: 18888 } });
|
||||
|
||||
expectCallField(mocks.auditGatewayServiceConfig, "expectedPort", 18888);
|
||||
const installOptions = requireRecord(
|
||||
callArg(mocks.install, 0, "install call"),
|
||||
"install options",
|
||||
expect(mocks.auditGatewayServiceConfig).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
expectedPort: 18888,
|
||||
}),
|
||||
);
|
||||
expect(mocks.install).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
programArguments: expect.arrayContaining(["18888"]),
|
||||
}),
|
||||
);
|
||||
expect(installOptions.programArguments).toContain("18888");
|
||||
});
|
||||
|
||||
it("repairs gateway services with embedded proxy environment values", async () => {
|
||||
@@ -523,14 +473,34 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair(cfg);
|
||||
|
||||
expectCallField(mocks.auditGatewayServiceConfig, "expectedGatewayToken", "env-token");
|
||||
expectCallConfigGatewayAuthToken(mocks.buildGatewayInstallPlan, "env-token");
|
||||
const replaceOptions = requireRecord(
|
||||
callArg(mocks.replaceConfigFile, 0, "replaceConfigFile call"),
|
||||
"replaceConfigFile options",
|
||||
expect(mocks.auditGatewayServiceConfig).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
expectedGatewayToken: "env-token",
|
||||
}),
|
||||
);
|
||||
expect(mocks.buildGatewayInstallPlan).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
config: expect.objectContaining({
|
||||
gateway: expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
token: "env-token",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(mocks.replaceConfigFile).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
nextConfig: expect.objectContaining({
|
||||
gateway: expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
token: "env-token",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
afterWrite: { mode: "auto" },
|
||||
}),
|
||||
);
|
||||
expectGatewayAuthToken(replaceOptions.nextConfig, "env-token");
|
||||
expect(replaceOptions.afterWrite).toEqual({ mode: "auto" });
|
||||
expect(mocks.stage).not.toHaveBeenCalled();
|
||||
expect(mocks.install).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
@@ -600,17 +570,18 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair({ gateway: {} });
|
||||
|
||||
const installPlanOptions = requireRecord(
|
||||
callArg(mocks.buildGatewayInstallPlan, 0, "buildGatewayInstallPlan call"),
|
||||
"buildGatewayInstallPlan options",
|
||||
expect(mocks.buildGatewayInstallPlan).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
env: expect.objectContaining({
|
||||
OPENCLAW_WRAPPER: wrapperPath,
|
||||
}),
|
||||
existingEnvironment: expect.objectContaining({
|
||||
OPENCLAW_WRAPPER: wrapperPath,
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expect(requireRecord(installPlanOptions.env, "install env").OPENCLAW_WRAPPER).toBe(wrapperPath);
|
||||
expect(
|
||||
requireRecord(installPlanOptions.existingEnvironment, "install existing environment")
|
||||
.OPENCLAW_WRAPPER,
|
||||
).toBe(wrapperPath);
|
||||
expectNoNoteContaining(
|
||||
"Gateway service entrypoint does not match the current install.",
|
||||
expect(mocks.note).not.toHaveBeenCalledWith(
|
||||
expect.stringContaining("Gateway service entrypoint does not match the current install."),
|
||||
"Gateway service config",
|
||||
);
|
||||
expect(mocks.note).toHaveBeenCalledWith(
|
||||
@@ -817,8 +788,16 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair(cfg);
|
||||
|
||||
expectCallField(mocks.auditGatewayServiceConfig, "expectedGatewayToken", undefined);
|
||||
expectCallField(mocks.buildGatewayInstallPlan, "config", cfg);
|
||||
expect(mocks.auditGatewayServiceConfig).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
expectedGatewayToken: undefined,
|
||||
}),
|
||||
);
|
||||
expect(mocks.buildGatewayInstallPlan).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
config: cfg,
|
||||
}),
|
||||
);
|
||||
expect(mocks.stage).not.toHaveBeenCalled();
|
||||
expect(mocks.install).toHaveBeenCalledTimes(1);
|
||||
});
|
||||
@@ -837,14 +816,34 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
|
||||
await runRepair(cfg);
|
||||
|
||||
expectCallField(mocks.auditGatewayServiceConfig, "expectedGatewayToken", undefined);
|
||||
const replaceOptions = requireRecord(
|
||||
callArg(mocks.replaceConfigFile, 0, "replaceConfigFile call"),
|
||||
"replaceConfigFile options",
|
||||
expect(mocks.auditGatewayServiceConfig).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
expectedGatewayToken: undefined,
|
||||
}),
|
||||
);
|
||||
expect(mocks.replaceConfigFile).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
nextConfig: expect.objectContaining({
|
||||
gateway: expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
token: "stale-token",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
afterWrite: { mode: "auto" },
|
||||
}),
|
||||
);
|
||||
expect(mocks.buildGatewayInstallPlan).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
config: expect.objectContaining({
|
||||
gateway: expect.objectContaining({
|
||||
auth: expect.objectContaining({
|
||||
token: "stale-token",
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
}),
|
||||
);
|
||||
expectGatewayAuthToken(replaceOptions.nextConfig, "stale-token");
|
||||
expect(replaceOptions.afterWrite).toEqual({ mode: "auto" });
|
||||
expectCallConfigGatewayAuthToken(mocks.buildGatewayInstallPlan, "stale-token");
|
||||
expect(mocks.stage).not.toHaveBeenCalled();
|
||||
expect(mocks.install).toHaveBeenCalledTimes(1);
|
||||
},
|
||||
@@ -922,7 +921,11 @@ describe("maybeRepairGatewayServiceConfig", () => {
|
||||
await runRepair(cfg);
|
||||
|
||||
expect(mocks.replaceConfigFile).not.toHaveBeenCalled();
|
||||
expectCallField(mocks.buildGatewayInstallPlan, "config", cfg);
|
||||
expect(mocks.buildGatewayInstallPlan).toHaveBeenCalledWith(
|
||||
expect.objectContaining({
|
||||
config: cfg,
|
||||
}),
|
||||
);
|
||||
expect(mocks.stage).not.toHaveBeenCalled();
|
||||
},
|
||||
);
|
||||
|
||||
@@ -409,18 +409,6 @@ export const FIELD_HELP: Record<string, string> = {
|
||||
"Experimental built-in tool flags. Keep these off by default and enable only when you are intentionally testing a preview surface.",
|
||||
"tools.experimental.planTool":
|
||||
"Enable the experimental structured `update_plan` tool for non-trivial multi-step work tracking. Leave this off unless you explicitly want the tool outside strict-agentic embedded Pi runs.",
|
||||
"tools.toolSearch":
|
||||
"Compact large OpenClaw, MCP, and client tool catalogs behind one search/call surface. Set to true for the default code bridge or use the object form to choose the structured fallback.",
|
||||
"tools.toolSearch.enabled":
|
||||
"Enables Tool Search. When on, OpenClaw hides large tool catalogs behind `tool_search_code` or structured search/describe/call tools during PI runs.",
|
||||
"tools.toolSearch.mode":
|
||||
'Choose the model-facing surface: "code" exposes `tool_search_code`; "tools" exposes structured search/describe/call fallback tools.',
|
||||
"tools.toolSearch.codeTimeoutMs":
|
||||
"Maximum milliseconds for one `tool_search_code` execution. Runtime clamps values to the supported 1s..60s range.",
|
||||
"tools.toolSearch.searchDefaultLimit":
|
||||
"Default number of Tool Search results returned when the model omits a limit. Runtime clamps this to `maxSearchLimit`.",
|
||||
"tools.toolSearch.maxSearchLimit":
|
||||
"Maximum number of Tool Search results a model can request. Runtime clamps values to the supported 1..50 range.",
|
||||
"tools.elevated":
|
||||
"Elevated tool access controls for privileged command surfaces that should only be reachable from trusted senders. Keep disabled unless operator workflows explicitly require elevated actions.",
|
||||
"tools.elevated.enabled":
|
||||
|
||||
@@ -238,12 +238,6 @@ export const FIELD_LABELS: Record<string, string> = {
|
||||
"tools.agentToAgent.allow": "Agent-to-Agent Target Allowlist",
|
||||
"tools.experimental": "Experimental Tools",
|
||||
"tools.experimental.planTool": "Enable Structured Plan Tool",
|
||||
"tools.toolSearch": "Tool Search",
|
||||
"tools.toolSearch.enabled": "Enable Tool Search",
|
||||
"tools.toolSearch.mode": "Tool Search Surface",
|
||||
"tools.toolSearch.codeTimeoutMs": "Tool Search Code Timeout",
|
||||
"tools.toolSearch.searchDefaultLimit": "Tool Search Default Results",
|
||||
"tools.toolSearch.maxSearchLimit": "Tool Search Max Results",
|
||||
"tools.elevated": "Elevated Tool Access",
|
||||
"tools.elevated.enabled": "Enable Elevated Tool Access",
|
||||
"tools.elevated.allowFrom": "Elevated Tool Allow Rules",
|
||||
|
||||
@@ -340,35 +340,6 @@ describe("config schema", () => {
|
||||
expect(parsed?.experimental?.planTool).toBe(true);
|
||||
});
|
||||
|
||||
it("accepts simplified Tool Search config in the runtime zod schema", () => {
|
||||
expect(ToolsSchema.parse({ toolSearch: true })?.toolSearch).toBe(true);
|
||||
expect(
|
||||
ToolsSchema.parse({
|
||||
toolSearch: {
|
||||
enabled: true,
|
||||
mode: "tools",
|
||||
codeTimeoutMs: 5000,
|
||||
searchDefaultLimit: 4,
|
||||
maxSearchLimit: 12,
|
||||
},
|
||||
})?.toolSearch,
|
||||
).toEqual({
|
||||
enabled: true,
|
||||
mode: "tools",
|
||||
codeTimeoutMs: 5000,
|
||||
searchDefaultLimit: 4,
|
||||
maxSearchLimit: 12,
|
||||
});
|
||||
expect(
|
||||
ToolsSchema.safeParse({
|
||||
toolSearch: {
|
||||
enabled: true,
|
||||
mode: "both",
|
||||
},
|
||||
}).success,
|
||||
).toBe(false);
|
||||
});
|
||||
|
||||
it("accepts web fetch maxResponseBytes in the runtime zod schema", () => {
|
||||
const parsed = ToolsSchema.parse({
|
||||
web: {
|
||||
|
||||
@@ -10,7 +10,6 @@ import { loadSessionStore } from "./store-load.js";
|
||||
import {
|
||||
resolveAgentSessionStoreTargetsSync,
|
||||
resolveAllAgentSessionStoreTargetsSync,
|
||||
resolveSessionStoreTargets,
|
||||
} from "./targets.js";
|
||||
import type { SessionEntry } from "./types.js";
|
||||
|
||||
@@ -60,7 +59,7 @@ function mergeSessionEntryIntoCombined(params: {
|
||||
|
||||
export function loadCombinedSessionStoreForGateway(
|
||||
cfg: OpenClawConfig,
|
||||
opts: { agentId?: string; configuredAgentsOnly?: boolean } = {},
|
||||
opts: { agentId?: string } = {},
|
||||
): {
|
||||
storePath: string;
|
||||
store: Record<string, SessionEntry>;
|
||||
@@ -94,9 +93,7 @@ export function loadCombinedSessionStoreForGateway(
|
||||
: undefined;
|
||||
const targets = requestedAgentId
|
||||
? resolveAgentSessionStoreTargetsSync(cfg, requestedAgentId)
|
||||
: opts.configuredAgentsOnly === true
|
||||
? resolveSessionStoreTargets(cfg, { allAgents: true })
|
||||
: resolveAllAgentSessionStoreTargetsSync(cfg);
|
||||
: resolveAllAgentSessionStoreTargetsSync(cfg);
|
||||
const combined: Record<string, SessionEntry> = {};
|
||||
for (const target of targets) {
|
||||
const agentId = target.agentId;
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user