docs(changelog): note patched Hono runtime

Update the global Hono override and published shrinkwraps to 4.12.25 so release packages avoid the current high-severity CORS advisory.

.agents/skills/crabbox/SKILL.md

@@ -54,6 +54,13 @@ pnpm crabbox:run -- --help | sed -n '1,120p'
 - For broad OpenClaw maintainer `pnpm` gates, prefer the repo wrapper with
   `--provider blacksmith-testbox` or the repo Testbox helpers when the standing
   Testbox policy applies.
+- Cold Testbox acquisition and hydration often take tens of seconds. When broad
+  remote proof is likely, immediately start
+  `node scripts/crabbox-wrapper.mjs warmup --provider blacksmith-testbox --keep --timing-json`
+  in a background command session while inspecting, editing, and running
+  focused local tests. Poll later, reuse the returned `tbx_...` with
+  `--provider blacksmith-testbox --id <tbx_id>`, and stop it before handoff.
+  Do not warm speculatively when remote proof is unlikely.
 - Always report the actual provider and id. `cbx_...` means AWS Crabbox;
   `tbx_...` means Blacksmith Testbox through Crabbox. If the output only says
   `blacksmith testbox list`, use `blacksmith testbox list --all` before

.agents/skills/release-openclaw-ci/SKILL.md

@@ -16,6 +16,10 @@ Use this with `$release-openclaw-maintainer` and `$openclaw-testing` when a rele
 - Watch one parent run plus compact child summaries. Avoid broad `gh run view` polling loops; REST quota is easy to burn.
 - Fetch logs only for failed or currently-blocking jobs. If quota is low, stop polling and wait for reset.
 - Treat live-provider flakes separately from code failures: prove key validity, provider HTTP status, retry evidence, and exact failing lane before editing code.
+- A model-list response proves authentication, not billing or inference
+  entitlement. Mandatory live providers must pass a real completion probe
+  before release dispatch. Fix the credential first; do not add an alternate
+  auth path merely to bypass a failed release credential.
 - Full Release Validation parent monitors fail fast: once a required child job
   fails, the parent cancels the remaining child matrix and prints the failed
   job summary. Inspect that first red job instead of waiting for unrelated
@@ -36,6 +40,8 @@ git rev-parse HEAD
 preflight. Inject those exact targeted keys first, then run the verifier; use
 ambient env only when it was already intentionally injected for this release.
 The script prints only provider status and HTTP class, never tokens.
+The Anthropic check performs a tiny message completion so exhausted or
+non-billable credentials fail before the expensive release matrix.
 
 ## Dispatch
 
@@ -106,7 +112,10 @@ Stop watchers before ending the turn or switching strategy.
      --jq '.jobs[] | select(.conclusion=="failure" or .conclusion=="timed_out" or .conclusion=="cancelled") | [.databaseId,.name,.conclusion,.url] | @tsv'
    ```
 3. Fetch one failed job log. If rate-limited, note reset time and avoid more REST calls.
-4. For secret-looking failures, validate the provider endpoint from the same secret source before editing code.
+4. For secret-looking failures, validate a real completion from the same secret source before editing code. A successful model-list request is insufficient.
+   Claude CLI subscription credentials are a separate native auth path; prove
+   them in a clean-home CLI probe, never as a substitute for a required
+   Anthropic API-key lane.
 5. For live-cache failures, inspect whether it is missing/invalid key, empty text, provider refusal, timeout, or baseline miss. Do not weaken release gates without clear provider evidence.
 6. Fix narrowly, run local/changed proof, commit, push, rerun the smallest matching group.
 

.agents/skills/release-openclaw-ci/scripts/verify-provider-secrets.mjs

@@ -1,17 +1,22 @@
 #!/usr/bin/env node
 /**
- * Release preflight helper that verifies required provider API keys can reach
- * their model-list endpoints without printing secret values.
+ * Release preflight helper that verifies required provider API keys without
+ * printing secret values. Anthropic must complete a prompt because model-list
+ * access does not prove billing or inference entitlement.
  */
 import process from "node:process";
 
 const args = new Map();
 for (let index = 2; index < process.argv.length; index += 1) {
   const arg = process.argv[index];
-  if (!arg.startsWith("--")) continue;
+  if (!arg.startsWith("--")) {
+    continue;
+  }
   const [key, inlineValue] = arg.slice(2).split("=", 2);
   const value = inlineValue ?? process.argv[index + 1];
-  if (inlineValue === undefined) index += 1;
+  if (inlineValue === undefined) {
+    index += 1;
+  }
   args.set(key, value);
 }
 
@@ -28,7 +33,9 @@ const timeoutMs = Number(args.get("timeout-ms") ?? 10_000);
 function envFirst(names) {
   for (const name of names) {
     const value = process.env[name]?.trim();
-    if (value) return { name, value };
+    if (value) {
+      return { name, value };
+    }
   }
   return undefined;
 }
@@ -44,13 +51,19 @@ async function checkProvider(id, config) {
   try {
     const headers = config.headers(secret.value);
     const response = await fetch(config.url, {
+      body: config.body,
       headers,
+      method: config.method,
       signal: controller.signal,
     });
+    const responseBody = config.validateResponse
+      ? await response.json().catch(() => undefined)
+      : undefined;
+    const ok = response.ok && (!config.validateResponse || config.validateResponse(responseBody));
     return {
       id,
-      ok: response.ok,
-      status: response.ok ? "ok" : `http_${response.status}`,
+      ok,
+      status: response.ok ? (ok ? "ok" : "invalid_response") : `http_${response.status}`,
       env: secret.name,
     };
   } catch (error) {
@@ -73,11 +86,21 @@ const providers = {
   },
   anthropic: {
     env: ["ANTHROPIC_API_KEY", "ANTHROPIC_API_TOKEN"],
-    url: "https://api.anthropic.com/v1/models",
+    url: "https://api.anthropic.com/v1/messages",
+    method: "POST",
+    body: JSON.stringify({
+      max_tokens: 8,
+      messages: [{ role: "user", content: "Reply with OK." }],
+      model: "claude-haiku-4-5",
+    }),
     headers: (token) => ({
       "anthropic-version": "2023-06-01",
+      "content-type": "application/json",
       "x-api-key": token,
     }),
+    validateResponse: (body) =>
+      Array.isArray(body?.content) &&
+      body.content.some((part) => typeof part?.text === "string" && part.text.trim()),
   },
   fireworks: {
     env: ["FIREWORKS_API_KEY"],
@@ -108,7 +131,9 @@ let failed = false;
 for (const result of results) {
   const requiredLabel = required.has(result.id) ? "required" : "optional";
   console.log(`${result.id}: ${result.status} env=${result.env} ${requiredLabel}`);
-  if (required.has(result.id) && !result.ok) failed = true;
+  if (required.has(result.id) && !result.ok) {
+    failed = true;
+  }
 }
 
 if (failed) {

.agents/skills/release-openclaw-maintainer/SKILL.md

@@ -150,9 +150,21 @@ Use this skill for release and publish-time workflow. Load `$release-private` if
 - Stable Windows Hub release closeout requires the signed
   `OpenClawCompanion-Setup-x64.exe`, `OpenClawCompanion-Setup-arm64.exe`, and
   `OpenClawCompanion-SHA256SUMS.txt` assets on the canonical
-  `openclaw/openclaw` GitHub Release. Use the public `Windows Node Release`
-  workflow after the matching `openclaw/openclaw-windows-node` release exists;
-  it verifies Authenticode signatures on Windows before uploading assets.
+  `openclaw/openclaw` GitHub Release. Pass the exact signed
+  `openclaw/openclaw-windows-node` release tag as `windows_node_tag` to
+  `OpenClaw Release Publish`, together with the candidate-approved
+  `windows_node_installer_digests` map; it prevalidates the published source
+  release and required installers against that map before any publish child,
+  dispatches the public `Windows Node Release` workflow while the OpenClaw
+  release is still a draft, carries those pinned source asset digests
+  unchanged, verifies the expected OpenClaw Foundation Authenticode signer on
+  Windows, re-downloads and checksum-verifies the promoted asset contract, and
+  blocks publication until the canonical asset contract is present. Use direct
+  `Windows Node Release` dispatch only for recovery, always with an exact tag,
+  never `latest`, and the explicit `expected_installer_digests` JSON map from
+  the approved source release. Recovery rejects unexpected
+  `OpenClawCompanion-*` target asset names, then replaces the expected contract
+  assets with the pinned source bytes.
 - Website Windows Hub download links should target exact canonical
   `openclaw/openclaw/releases/download/vYYYY.M.PATCH/...` assets for the current
   stable release, or `releases/latest/download/...` only after verifying the
@@ -309,6 +321,7 @@ Upgrade with the beta channel.
 Before tagging or publishing, run:
 
 ```bash
+pnpm release:fast-pretag-check
 pnpm check:architecture
 pnpm build
 pnpm ui:build
@@ -317,6 +330,21 @@ pnpm release:check
 pnpm test:install:smoke
 ```
 
+- Treat `pnpm release:fast-pretag-check` as a hard packaging gate. Every
+  publishable plugin must have a non-empty package-root `README.md`, build its
+  package-local runtime, and pass the npm and ClawHub release metadata checks
+  before a tag or publish workflow can start. Do not defer README, entrypoint,
+  or packed-artifact failures to postpublish verification.
+- Before tagging, require green CI for the exact release-candidate SHA, not an
+  earlier branch SHA. Heal every related red CI, release-check, packaging, or
+  root-Dockerfile lane on the release branch, forward-port the fix to `main`,
+  and rerun the affected exact-SHA gates. Never waive a red Docker lane because
+  npm preflight passed.
+- Root Dockerfile proof is mandatory before every beta and stable tag. Run the
+  release `install-smoke` group or equivalent root Dockerfile build for the
+  exact candidate SHA and require it to pass. The tag-triggered Docker Release
+  workflow is post-tag publishing, not the first valid proof that the root
+  Dockerfile can build.
 - Before tagging, diff publishable plugin package manifests against the last
   reachable stable/beta release tag. For every newly publishable package
   (`openclaw.release.publishToNpm: true` or `publishToClawHub: true`) whose
@@ -524,6 +552,16 @@ node --import tsx scripts/openclaw-npm-postpublish-verify.ts <published-version>
 - `preflight_only=true` on the npm workflow is also the right way to validate an
   existing tag after publish; it should keep running the build checks even when
   the npm version is already published.
+- npm registry metadata is eventually consistent immediately after trusted
+  publishing. Keep postpublish `npm view` checks on bounded `--prefer-online`
+  retries, and carry that verified tarball/integrity metadata into later proof
+  steps instead of reading the registry again. If the OpenClaw npm child
+  succeeded but the parent publish workflow failed on an immediate exact-version
+  `E404`, verify the exact version with a cache-bypassed registry read, run the
+  standalone postpublish verifier and the full beta verifier with the original
+  successful child run IDs, then finalize the draft, dependency evidence asset,
+  and release proof manually. Never rerun the publish workflow for that
+  already-published version.
 - npm validation-only preflight may still be dispatched from ordinary branches
   when testing workflow changes before merge. Release checks and real publish
   use only `main` or `release/YYYY.M.PATCH`.
@@ -632,9 +670,10 @@ node --import tsx scripts/openclaw-npm-postpublish-verify.ts <published-version>
    off, live OpenAI off, and regression failure off. Let it run in parallel
    with preflight and validation work.
 10. Run the fast local beta preflight from the release branch before any npm
-    preflight or publish. Keep expensive Docker, Parallels, and published-package
-    install/update lanes for after the beta is live unless the operator asks to
-    run them before beta publication.
+    preflight or publish. Require exact-SHA CI and root Dockerfile install-smoke
+    to be green before tagging. Keep the remaining expensive Docker, Parallels,
+    and published-package install/update lanes for after the beta is live unless
+    the operator asks to run them before beta publication.
 11. For beta releases, skip mac app build/sign/notarize unless beta scope or a
     release blocker specifically requires it. For stable releases, include the
     mac app, signing, notarization, and appcast path.
@@ -675,20 +714,29 @@ node --import tsx scripts/openclaw-npm-postpublish-verify.ts <published-version>
     where npm did not publish the beta version, delete/recreate the same beta
     tag and any accidental draft/incomplete prerelease at the fixed commit
     instead of skipping a prerelease number.
-22. Start `.github/workflows/openclaw-npm-release.yml` from the same branch with
+22. Start `.github/workflows/openclaw-release-publish.yml` from the same branch with
     the same tag for the real publish, choose `npm_dist_tag` (`beta` default,
     `latest` only when you intentionally want direct stable publish), keep it
     the same as the preflight run, and pass the successful npm
-    `preflight_run_id`.
+    `preflight_run_id` plus the successful `full_release_validation_run_id`.
+    For stable publish, also pass the exact non-prerelease
+    `openclaw/openclaw-windows-node` tag as `windows_node_tag` and its
+    candidate-approved installer digest map as `windows_node_installer_digests`.
 23. Wait for `npm-release` approval from `@openclaw/openclaw-release-managers`.
 24. Wait for the real publish workflow to run postpublish verification,
     create or update the GitHub release as a draft, upload dependency evidence,
+    promote and verify the required Windows Hub assets for stable releases,
     append release verification proof, and only then undraft/publish it. If a
-    waited plugin publish fails after OpenClaw npm succeeds, the workflow keeps
-    the release draft with OpenClaw npm evidence and exits red; do not undraft
-    until the plugin publish gap is repaired. The standalone verifier command
-    remains the recovery probe:
+    waited plugin publish or Windows Hub promotion fails after OpenClaw npm
+    succeeds, the workflow keeps the release draft with OpenClaw npm evidence
+    and exits red; do not undraft until the gap is repaired. The standalone
+    verifier command remains the first recovery probe:
     `node --import tsx scripts/openclaw-npm-postpublish-verify.ts <published-version>`.
+    For a failed postpublish parent after successful publish children, also run
+    `pnpm release:verify-beta -- <published-version> ... --skip-github-release`
+    with the original child run IDs and an evidence output path before manually
+    recreating the workflow's draft, dependency evidence asset, proof section,
+    and publish step.
 25. Run the post-published beta verification roster. First scan current `main`
     for critical fixes that landed after the release branch cut; backport only
     important low-risk fixes before starting expensive lanes, or increment to

.github/codeql/codeql-process-exec-boundary-critical-security.yml

@@ -1,61 +0,0 @@
-name: openclaw-codeql-process-exec-boundary-critical-security
-
-disable-default-queries: true
-
-queries:
-  - uses: security-extended
-
-query-filters:
-  - include:
-      precision:
-        - high
-        - very-high
-      tags contain: security
-      security-severity: /([7-9]|10)\.(\d)+/
-
-paths:
-  - src/process
-  - src/tui/tui-local-shell.ts
-  - src/tui/tui.ts
-  - src/plugin-sdk/windows-spawn.ts
-  - packages/agent-core/src/harness/env
-  - packages/memory-host-sdk/src/host
-  - extensions/acpx/src
-  - extensions/bonjour/src/advertiser.ts
-  - extensions/browser/src/browser/chrome-mcp.ts
-  - extensions/browser/src/browser/chrome.executables.ts
-  - extensions/browser/src/browser/chrome.ts
-  - extensions/codex/src/app-server/sandbox-exec-server
-  - extensions/codex/src/app-server/transport-stdio.ts
-  - extensions/codex/src/node-cli-sessions.ts
-  - extensions/codex-supervisor/src/json-rpc-client.ts
-  - extensions/file-transfer/src
-  - extensions/google-meet/src
-  - extensions/imessage/src
-  - extensions/memory-core/src/memory/qmd-manager.ts
-  - extensions/memory-wiki/src/obsidian.ts
-  - extensions/microsoft-foundry/cli.ts
-  - extensions/ollama/src/wsl2-crash-loop-check.ts
-  - extensions/qa-lab/src
-  - extensions/signal/src/daemon.ts
-  - extensions/tts-local-cli/speech-provider.ts
-  - extensions/voice-call/src
-  - scripts
-
-paths-ignore:
-  - "**/node_modules"
-  - "**/coverage"
-  - "**/*.generated.ts"
-  - "**/*.bundle.js"
-  - "**/*-runtime.js"
-  - "**/*.test.ts"
-  - "**/*.test.tsx"
-  - "**/*.spec.ts"
-  - "**/*.spec.tsx"
-  - "**/*.e2e.test.ts"
-  - "**/*.e2e.test.tsx"
-  - "**/*test-support*"
-  - "**/*test-helper*"
-  - "**/*mock*"
-  - "**/*fixture*"
-  - "**/*bench*"

.github/workflows/codeql.yml

@@ -17,28 +17,7 @@ on:
       - ".github/actions/**"
       - ".github/codeql/**"
       - ".github/workflows/**"
-      - "extensions/acpx/src/**"
-      - "extensions/bonjour/src/advertiser.ts"
-      - "extensions/browser/src/browser/chrome-mcp.ts"
-      - "extensions/browser/src/browser/chrome.executables.ts"
-      - "extensions/browser/src/browser/chrome.ts"
-      - "extensions/codex/src/app-server/sandbox-exec-server/**"
-      - "extensions/codex/src/app-server/transport-stdio.ts"
-      - "extensions/codex/src/node-cli-sessions.ts"
-      - "extensions/codex-supervisor/src/json-rpc-client.ts"
-      - "extensions/file-transfer/src/**"
-      - "extensions/google-meet/src/**"
-      - "extensions/imessage/src/**"
-      - "extensions/memory-core/src/memory/qmd-manager.ts"
-      - "extensions/memory-wiki/src/obsidian.ts"
-      - "extensions/microsoft-foundry/cli.ts"
-      - "extensions/ollama/src/wsl2-crash-loop-check.ts"
-      - "extensions/qa-lab/src/**"
-      - "extensions/signal/src/daemon.ts"
-      - "extensions/tts-local-cli/speech-provider.ts"
-      - "extensions/voice-call/src/**"
       - "packages/**"
-      - "scripts/**"
       - "src/**"
   push:
     branches:
@@ -47,28 +26,7 @@ on:
       - ".github/actions/**"
       - ".github/codeql/**"
       - ".github/workflows/**"
-      - "extensions/acpx/src/**"
-      - "extensions/bonjour/src/advertiser.ts"
-      - "extensions/browser/src/browser/chrome-mcp.ts"
-      - "extensions/browser/src/browser/chrome.executables.ts"
-      - "extensions/browser/src/browser/chrome.ts"
-      - "extensions/codex/src/app-server/sandbox-exec-server/**"
-      - "extensions/codex/src/app-server/transport-stdio.ts"
-      - "extensions/codex/src/node-cli-sessions.ts"
-      - "extensions/codex-supervisor/src/json-rpc-client.ts"
-      - "extensions/file-transfer/src/**"
-      - "extensions/google-meet/src/**"
-      - "extensions/imessage/src/**"
-      - "extensions/memory-core/src/memory/qmd-manager.ts"
-      - "extensions/memory-wiki/src/obsidian.ts"
-      - "extensions/microsoft-foundry/cli.ts"
-      - "extensions/ollama/src/wsl2-crash-loop-check.ts"
-      - "extensions/qa-lab/src/**"
-      - "extensions/signal/src/daemon.ts"
-      - "extensions/tts-local-cli/speech-provider.ts"
-      - "extensions/voice-call/src/**"
       - "packages/**"
-      - "scripts/**"
       - "src/**"
   schedule:
     - cron: "0 6 * * *"
@@ -115,11 +73,6 @@ jobs:
             runs_on: blacksmith-4vcpu-ubuntu-2404
             timeout_minutes: 25
             config_file: ./.github/codeql/codeql-mcp-process-tool-boundary-critical-security.yml
-          - language: javascript-typescript
-            category: process-exec-boundary
-            runs_on: blacksmith-4vcpu-ubuntu-2404
-            timeout_minutes: 25
-            config_file: ./.github/codeql/codeql-process-exec-boundary-critical-security.yml
           - language: javascript-typescript
             category: plugin-trust-boundary
             runs_on: blacksmith-4vcpu-ubuntu-2404

.github/workflows/full-release-validation.yml

@@ -783,7 +783,7 @@ jobs:
           fi
 
           args=(
-            -f ref="$TARGET_SHA"
+            -f ref="$TARGET_REF"
             -f expected_sha="$TARGET_SHA"
             -f provider="$PROVIDER"
             -f mode="$MODE"

.github/workflows/mantis-telegram-live.yml

@@ -379,7 +379,6 @@ jobs:
           OPENCLAW_QA_CONVEX_SECRET_CI: ${{ secrets.OPENCLAW_QA_CONVEX_SECRET_CI }}
           OPENCLAW_QA_CREDENTIAL_ACQUIRE_TIMEOUT_MS: "1800000"
           OPENCLAW_QA_REDACT_PUBLIC_METADATA: "1"
-          OPENCLAW_QA_TELEGRAM_CAPTURE_CONTENT: "1"
           CRABBOX_COORDINATOR: ${{ secrets.CRABBOX_COORDINATOR }}
           CRABBOX_COORDINATOR_TOKEN: ${{ secrets.CRABBOX_COORDINATOR_TOKEN }}
           OPENCLAW_QA_MANTIS_CRABBOX_COORDINATOR: ${{ secrets.OPENCLAW_QA_MANTIS_CRABBOX_COORDINATOR }}

.github/workflows/npm-telegram-beta-e2e.yml

@@ -220,7 +220,6 @@ jobs:
           OPENCLAW_QA_CONVEX_SECRET_CI: ${{ secrets.OPENCLAW_QA_CONVEX_SECRET_CI }}
           OPENCLAW_QA_CREDENTIAL_ACQUIRE_TIMEOUT_MS: "1800000"
           OPENCLAW_QA_REDACT_PUBLIC_METADATA: "1"
-          OPENCLAW_QA_TELEGRAM_CAPTURE_CONTENT: "1"
           INPUT_SCENARIO: ${{ inputs.scenario }}
           PACKAGE_ARTIFACT_NAME: ${{ inputs.package_artifact_name || '' }}
         run: |

.github/workflows/openclaw-live-and-e2e-checks-reusable.yml

@@ -420,6 +420,7 @@ jobs:
               add_suite live-cache
 
               add_profile_suite native-live-src-agents "stable full"
+              add_profile_suite native-live-src-agents-zai-coding "stable full"
               add_profile_suite native-live-src-gateway-core "beta minimum stable full"
               add_profile_suite native-live-src-gateway-profiles-anthropic "stable full"
               add_profile_suite native-live-src-gateway-profiles-anthropic-smoke "stable"
@@ -1956,6 +1957,12 @@ jobs:
             timeout_minutes: 60
             profile_env_only: false
             profiles: stable full
+          - suite_id: native-live-src-agents-zai-coding
+            label: Native live Z.AI Coding Plan
+            command: ZAI_CODING_LIVE_TEST=1 node .release-harness/scripts/test-live-shard.mjs native-live-src-agents-zai-coding
+            timeout_minutes: 15
+            profile_env_only: false
+            profiles: stable full
           - suite_id: native-live-src-gateway-core
             label: Native live gateway core
             command: OPENCLAW_LIVE_CODEX_HARNESS=1 OPENCLAW_LIVE_CODEX_HARNESS_AUTH=api-key node .release-harness/scripts/test-live-shard.mjs native-live-src-gateway-core
@@ -2215,7 +2222,11 @@ jobs:
           case "${{ matrix.suite_id }}" in
             live-cli-backend-docker)
               echo "OPENCLAW_LIVE_CLI_BACKEND_MODEL=claude-cli/claude-sonnet-4-6" >> "$GITHUB_ENV"
-              echo "OPENCLAW_LIVE_CLI_BACKEND_AUTH=api-key" >> "$GITHUB_ENV"
+              if [[ -n "${OPENCLAW_CLAUDE_CREDENTIALS_JSON:-}" || -n "${CLAUDE_CODE_OAUTH_TOKEN:-}" ]]; then
+                echo "OPENCLAW_LIVE_CLI_BACKEND_AUTH=subscription" >> "$GITHUB_ENV"
+              else
+                echo "OPENCLAW_LIVE_CLI_BACKEND_AUTH=api-key" >> "$GITHUB_ENV"
+              fi
               echo "OPENCLAW_LIVE_CLI_BACKEND_DEBUG=1" >> "$GITHUB_ENV"
               echo "OPENCLAW_CLI_BACKEND_LOG_OUTPUT=1" >> "$GITHUB_ENV"
               echo "OPENCLAW_TEST_CONSOLE=1" >> "$GITHUB_ENV"
@@ -2440,7 +2451,11 @@ jobs:
           case "${{ matrix.suite_id }}" in
             live-cli-backend-docker)
               echo "OPENCLAW_LIVE_CLI_BACKEND_MODEL=claude-cli/claude-sonnet-4-6" >> "$GITHUB_ENV"
-              echo "OPENCLAW_LIVE_CLI_BACKEND_AUTH=api-key" >> "$GITHUB_ENV"
+              if [[ -n "${OPENCLAW_CLAUDE_CREDENTIALS_JSON:-}" || -n "${CLAUDE_CODE_OAUTH_TOKEN:-}" ]]; then
+                echo "OPENCLAW_LIVE_CLI_BACKEND_AUTH=subscription" >> "$GITHUB_ENV"
+              else
+                echo "OPENCLAW_LIVE_CLI_BACKEND_AUTH=api-key" >> "$GITHUB_ENV"
+              fi
               echo "OPENCLAW_LIVE_CLI_BACKEND_DEBUG=1" >> "$GITHUB_ENV"
               echo "OPENCLAW_CLI_BACKEND_LOG_OUTPUT=1" >> "$GITHUB_ENV"
               echo "OPENCLAW_TEST_CONSOLE=1" >> "$GITHUB_ENV"

.github/workflows/openclaw-release-checks.yml

@@ -1181,7 +1181,7 @@ jobs:
   runtime_tool_coverage_release_checks:
     name: Enforce QA Lab runtime tool coverage
     needs: [resolve_target, qa_lab_runtime_parity_release_checks]
-    if: always() && contains(fromJSON('["all","qa","qa-parity"]'), needs.resolve_target.outputs.rerun_group)
+    if: contains(fromJSON('["all","qa","qa-parity"]'), needs.resolve_target.outputs.rerun_group)
     runs-on: ubuntu-24.04
     timeout-minutes: 15
     permissions:
@@ -1204,13 +1204,35 @@ jobs:
           node-version: ${{ env.NODE_VERSION }}
           install-bun: "true"
 
+      - name: Download runtime parity status
+        uses: actions/download-artifact@v8
+        with:
+          name: release-check-status-qa-runtime-parity-${{ needs.resolve_target.outputs.revision }}
+          path: .artifacts/release-check-status/
+
+      - name: Verify runtime parity producer status
+        id: verify_runtime_parity_status
+        shell: bash
+        run: |
+          set -euo pipefail
+          status_path=".artifacts/release-check-status/qa_lab_runtime_parity_release_checks.env"
+          status="$(sed -n 's/^status=//p' "$status_path" | tail -n 1)"
+          if [[ "$status" != "success" ]]; then
+            echo "Runtime parity producer status is ${status:-missing}; skipping coverage artifact consumer."
+            echo "ready=false" >> "$GITHUB_OUTPUT"
+            exit 0
+          fi
+          echo "ready=true" >> "$GITHUB_OUTPUT"
+
       - name: Download runtime parity artifacts
+        if: steps.verify_runtime_parity_status.outputs.ready == 'true'
         uses: actions/download-artifact@v8
         with:
           name: release-qa-runtime-parity-${{ needs.resolve_target.outputs.revision }}
           path: .artifacts/qa-e2e/
 
       - name: Enforce standard runtime tool coverage
+        if: steps.verify_runtime_parity_status.outputs.ready == 'true'
         run: |
           set -euo pipefail
           pnpm openclaw qa coverage \
@@ -1412,7 +1434,6 @@ jobs:
           OPENCLAW_QA_CONVEX_SECRET_CI: ${{ secrets.OPENCLAW_QA_CONVEX_SECRET_CI }}
           OPENCLAW_QA_CREDENTIAL_ACQUIRE_TIMEOUT_MS: "1800000"
           OPENCLAW_QA_REDACT_PUBLIC_METADATA: "1"
-          OPENCLAW_QA_TELEGRAM_CAPTURE_CONTENT: "1"
         run: |
           set -euo pipefail
 

.github/workflows/openclaw-release-publish.yml

@@ -15,6 +15,14 @@ on:
         description: Successful Full Release Validation run id for this tag/SHA, required when publish_openclaw_npm=true
         required: false
         type: string
+      windows_node_tag:
+        description: Exact openclaw-windows-node release tag, required for stable OpenClaw publish
+        required: false
+        type: string
+      windows_node_installer_digests:
+        description: Candidate-approved compact JSON map of Windows installer names to pinned sha256 digests
+        required: false
+        type: string
       npm_telegram_run_id:
         description: Optional successful NPM Telegram Beta E2E run id to include in final release evidence
         required: false
@@ -81,12 +89,15 @@ jobs:
     outputs:
       sha: ${{ steps.manifest.outputs.sha || steps.ref.outputs.sha }}
       preflight_artifact_name: ${{ steps.preflight_artifact.outputs.name }}
+      windows_node_installer_digests: ${{ steps.windows_source.outputs.installer_digests }}
     steps:
       - name: Validate inputs
         env:
           RELEASE_TAG: ${{ inputs.tag }}
           PREFLIGHT_RUN_ID: ${{ inputs.preflight_run_id }}
           FULL_RELEASE_VALIDATION_RUN_ID: ${{ inputs.full_release_validation_run_id }}
+          WINDOWS_NODE_TAG: ${{ inputs.windows_node_tag }}
+          WINDOWS_NODE_INSTALLER_DIGESTS: ${{ inputs.windows_node_installer_digests }}
           PUBLISH_OPENCLAW_NPM: ${{ inputs.publish_openclaw_npm && 'true' || 'false' }}
           PLUGIN_PUBLISH_SCOPE: ${{ inputs.plugin_publish_scope }}
           PLUGINS: ${{ inputs.plugins }}
@@ -115,6 +126,22 @@ jobs:
             echo "publish_openclaw_npm=true requires full_release_validation_run_id." >&2
             exit 1
           fi
+          stable_release=true
+          if [[ "${RELEASE_TAG}" == *"-alpha."* || "${RELEASE_TAG}" == *"-beta."* ]]; then
+            stable_release=false
+          fi
+          if [[ -n "${WINDOWS_NODE_TAG}" && ! "${WINDOWS_NODE_TAG}" =~ ^v[0-9]+\.[0-9]+\.[0-9]+([-.][0-9A-Za-z]+([.-][0-9A-Za-z]+)*)?$ ]]; then
+            echo "windows_node_tag must be an explicit openclaw-windows-node release tag, not latest: ${WINDOWS_NODE_TAG}" >&2
+            exit 1
+          fi
+          if [[ "${PUBLISH_OPENCLAW_NPM}" == "true" && "${stable_release}" == "true" && -z "${WINDOWS_NODE_TAG}" ]]; then
+            echo "Stable OpenClaw publish requires an explicit windows_node_tag." >&2
+            exit 1
+          fi
+          if [[ "${PUBLISH_OPENCLAW_NPM}" == "true" && "${stable_release}" == "true" && -z "${WINDOWS_NODE_INSTALLER_DIGESTS}" ]]; then
+            echo "Stable OpenClaw publish requires candidate-approved windows_node_installer_digests." >&2
+            exit 1
+          fi
           tideclaw_alpha_publish=false
           if [[ "${RELEASE_TAG}" == *"-alpha."* && "${RELEASE_NPM_DIST_TAG}" == "alpha" && "${WORKFLOW_REF}" =~ ^refs/heads/tideclaw/alpha/[0-9]{4}-[0-9]{2}-[0-9]{2}-[0-9]{4}Z$ ]]; then
             tideclaw_alpha_publish=true
@@ -143,6 +170,73 @@ jobs:
               ;;
           esac
 
+      - name: Validate stable Windows source release
+        id: windows_source
+        if: ${{ inputs.publish_openclaw_npm }}
+        env:
+          GH_TOKEN: ${{ github.token }}
+          RELEASE_TAG: ${{ inputs.tag }}
+          WINDOWS_NODE_TAG: ${{ inputs.windows_node_tag }}
+          APPROVED_INSTALLER_DIGESTS: ${{ inputs.windows_node_installer_digests }}
+        run: |
+          set -euo pipefail
+          if [[ "${RELEASE_TAG}" == *"-alpha."* || "${RELEASE_TAG}" == *"-beta."* ]]; then
+            exit 0
+          fi
+
+          source_json="$(gh release view "${WINDOWS_NODE_TAG}" \
+            --repo openclaw/openclaw-windows-node \
+            --json tagName,isDraft,isPrerelease,assets,url)"
+          if [[ "$(printf '%s' "${source_json}" | jq -r '.tagName')" != "${WINDOWS_NODE_TAG}" ]]; then
+            echo "Windows source release tag does not match ${WINDOWS_NODE_TAG}." >&2
+            exit 1
+          fi
+          if [[ "$(printf '%s' "${source_json}" | jq -r '.isDraft')" == "true" ]]; then
+            echo "Stable OpenClaw publish requires a published Windows source release." >&2
+            exit 1
+          fi
+          if [[ "$(printf '%s' "${source_json}" | jq -r '.isPrerelease')" == "true" ]]; then
+            echo "Stable OpenClaw publish requires a non-prerelease Windows source release." >&2
+            exit 1
+          fi
+
+          required_assets=(
+            "OpenClawCompanion-Setup-x64.exe"
+            "OpenClawCompanion-Setup-arm64.exe"
+          )
+          required_assets_json="$(printf '%s\n' "${required_assets[@]}" | jq -R . | jq -sc .)"
+          if ! approved_installer_digests="$(printf '%s' "${APPROVED_INSTALLER_DIGESTS}" | jq -ce --argjson names "${required_assets_json}" '
+            if type == "object" and
+              (keys | sort) == ($names | sort) and
+              all(.[]; type == "string" and test("^sha256:[a-f0-9]{64}$"))
+            then .
+            else error("invalid candidate-approved Windows installer digest map")
+            end
+          ')"; then
+            echo "windows_node_installer_digests must contain exactly the candidate-approved current installer asset contract." >&2
+            exit 1
+          fi
+          for asset_name in "${required_assets[@]}"; do
+            asset_matches="$(printf '%s' "${source_json}" | jq -c --arg name "${asset_name}" '[.assets[]? | select(.name == $name)]')"
+            asset_match_count="$(printf '%s' "${asset_matches}" | jq 'length')"
+            if [[ "${asset_match_count}" != "1" ]]; then
+              echo "Windows source release ${WINDOWS_NODE_TAG} must contain exactly one required asset ${asset_name}; found ${asset_match_count}." >&2
+              exit 1
+            fi
+            asset_digest="$(printf '%s' "${asset_matches}" | jq -r '.[0].digest // empty')"
+            if [[ ! "${asset_digest}" =~ ^sha256:[a-f0-9]{64}$ ]]; then
+              echo "Windows source release ${WINDOWS_NODE_TAG} asset ${asset_name} is missing its immutable SHA-256 digest." >&2
+              exit 1
+            fi
+            approved_digest="$(printf '%s' "${approved_installer_digests}" | jq -r --arg name "${asset_name}" '.[$name]')"
+            if [[ "${asset_digest}" != "${approved_digest}" ]]; then
+              echo "Windows source release ${WINDOWS_NODE_TAG} asset ${asset_name} no longer matches its candidate-approved digest." >&2
+              exit 1
+            fi
+          done
+          echo "installer_digests=${approved_installer_digests}" >> "$GITHUB_OUTPUT"
+          echo "- Windows Node source release: prevalidated \`${WINDOWS_NODE_TAG}\`" >> "$GITHUB_STEP_SUMMARY"
+
       - name: Download OpenClaw npm preflight manifest
         id: preflight_artifact
         if: ${{ inputs.publish_openclaw_npm }}
@@ -337,6 +431,7 @@ jobs:
           TARGET_SHA: ${{ steps.manifest.outputs.sha || steps.ref.outputs.sha }}
           RELEASE_PROFILE: ${{ steps.full_manifest.outputs.release_profile || inputs.release_profile }}
           FULL_RELEASE_VALIDATION_RUN_ID: ${{ inputs.full_release_validation_run_id }}
+          WINDOWS_NODE_TAG: ${{ inputs.windows_node_tag }}
         run: |
           {
             echo "### Release target"
@@ -347,13 +442,16 @@ jobs:
             if [[ -n "${FULL_RELEASE_VALIDATION_RUN_ID// }" ]]; then
               echo "- Full release validation: \`${FULL_RELEASE_VALIDATION_RUN_ID}\`"
             fi
+            if [[ -n "${WINDOWS_NODE_TAG// }" ]]; then
+              echo "- Windows Node source release: \`${WINDOWS_NODE_TAG}\`"
+            fi
           } >> "$GITHUB_STEP_SUMMARY"
 
   publish:
     name: Publish plugins, then OpenClaw
     needs: [resolve_release_target]
     runs-on: ubuntu-latest
-    timeout-minutes: 60
+    timeout-minutes: 120
     environment: npm-release
     steps:
       - name: Checkout release SHA
@@ -383,10 +481,16 @@ jobs:
           WAIT_FOR_CLAWHUB: ${{ inputs.wait_for_clawhub && 'true' || 'false' }}
           PREFLIGHT_ARTIFACT_NAME: ${{ needs.resolve_release_target.outputs.preflight_artifact_name }}
           NPM_TELEGRAM_RUN_ID: ${{ inputs.npm_telegram_run_id }}
+          WINDOWS_NODE_TAG: ${{ inputs.windows_node_tag }}
+          WINDOWS_NODE_INSTALLER_DIGESTS: ${{ needs.resolve_release_target.outputs.windows_node_installer_digests }}
           POSTPUBLISH_EVIDENCE_DIR: ${{ runner.temp }}/openclaw-release-postpublish-evidence
         run: |
           set -euo pipefail
 
+          is_stable_release() {
+            [[ "${RELEASE_TAG}" != *"-alpha."* && "${RELEASE_TAG}" != *"-beta."* ]]
+          }
+
           dispatch_workflow_at_ref() {
             local workflow_ref="$1"
             shift
@@ -836,10 +940,105 @@ jobs:
           }
 
           publish_github_release() {
+            if is_stable_release; then
+              verify_windows_release_asset_contract
+            fi
             gh release edit "${RELEASE_TAG}" --repo "$GITHUB_REPOSITORY" --draft=false
             echo "- GitHub release: https://github.com/${GITHUB_REPOSITORY}/releases/tag/${RELEASE_TAG}" >> "$GITHUB_STEP_SUMMARY"
           }
 
+          verify_windows_release_asset_contract() {
+            local actual_companion_assets actual_digest asset_name expected_companion_assets expected_digest expected_hash expected_installer_names manifest_dir manifest_json manifest_path release_json
+            # Add future promoted installer names, such as MSIX x64/ARM64, here.
+            local -a installer_assets=(
+              "OpenClawCompanion-Setup-x64.exe"
+              "OpenClawCompanion-Setup-arm64.exe"
+            )
+            local -a required_assets=(
+              "${installer_assets[@]}"
+              "OpenClawCompanion-SHA256SUMS.txt"
+            )
+
+            release_json="$(gh release view "${RELEASE_TAG}" --repo "$GITHUB_REPOSITORY" --json assets,url)"
+            expected_companion_assets="$(printf '%s\n' "${required_assets[@]}" | jq -R . | jq -sc 'sort')"
+            actual_companion_assets="$(printf '%s' "${release_json}" | jq -c '
+              [.assets[]? | select(.name | startswith("OpenClawCompanion-")) | .name] | sort
+            ')"
+            if [[ "${actual_companion_assets}" != "${expected_companion_assets}" ]]; then
+              echo "Stable release OpenClawCompanion asset names do not exactly match the current contract." >&2
+              return 1
+            fi
+            for asset_name in "${required_assets[@]}"; do
+              if ! printf '%s' "${release_json}" | jq -e --arg name "${asset_name}" 'any(.assets[]?; .name == $name)' >/dev/null; then
+                echo "Stable release is missing required Windows asset ${asset_name}." >&2
+                return 1
+              fi
+            done
+
+            manifest_dir="${RUNNER_TEMP}/openclaw-windows-release-contract"
+            manifest_path="${manifest_dir}/OpenClawCompanion-SHA256SUMS.txt"
+            rm -rf "${manifest_dir}"
+            mkdir -p "${manifest_dir}"
+            gh release download "${RELEASE_TAG}" \
+              --repo "$GITHUB_REPOSITORY" \
+              --pattern "OpenClawCompanion-SHA256SUMS.txt" \
+              --dir "${manifest_dir}"
+            if ! manifest_json="$(jq -Rsc '
+              split("\n") as $lines |
+              (if $lines[-1] == "" then $lines[0:-1] else $lines end) |
+              map(sub("\r$"; "")) |
+              if all(.[]; test("^(?<hash>[a-f0-9]{64})  (?<name>[^/\\\\]+)$"))
+              then map(capture("^(?<hash>[a-f0-9]{64})  (?<name>[^/\\\\]+)$"))
+              else error("malformed Windows checksum manifest entry")
+              end
+            ' "${manifest_path}")"; then
+              echo "Stable release Windows checksum manifest contains malformed entries." >&2
+              return 1
+            fi
+            expected_installer_names="$(printf '%s\n' "${installer_assets[@]}" | jq -R . | jq -sc 'sort')"
+            if ! printf '%s' "${manifest_json}" | jq -e --argjson expected "${expected_installer_names}" '
+              length == ($expected | length) and
+              ([.[].name] | sort) == $expected and
+              ([.[].name] | unique | length) == length
+            ' >/dev/null; then
+              echo "Stable release Windows checksum manifest does not exactly match the installer asset contract." >&2
+              return 1
+            fi
+            for asset_name in "${installer_assets[@]}"; do
+              expected_digest="$(printf '%s' "${WINDOWS_NODE_INSTALLER_DIGESTS}" | jq -r --arg name "${asset_name}" '.[$name] // empty')"
+              actual_digest="$(printf '%s' "${release_json}" | jq -r --arg name "${asset_name}" '.assets[]? | select(.name == $name) | .digest // empty')"
+              if [[ -z "${expected_digest}" || "${actual_digest}" != "${expected_digest}" ]]; then
+                echo "Stable release Windows asset ${asset_name} does not match its pinned digest." >&2
+                return 1
+              fi
+              expected_hash="${expected_digest#sha256:}"
+              if ! printf '%s' "${manifest_json}" | jq -e --arg name "${asset_name}" --arg hash "${expected_hash}" '
+                any(.[]; .name == $name and .hash == $hash)
+              ' >/dev/null; then
+                echo "Stable release Windows checksum manifest does not match pinned digest for ${asset_name}." >&2
+                return 1
+              fi
+            done
+            echo "- Windows Hub asset contract: verified" >> "$GITHUB_STEP_SUMMARY"
+          }
+
+          promote_windows_release_assets() {
+            if ! is_stable_release; then
+              return 0
+            fi
+            if [[ -z "${WINDOWS_NODE_INSTALLER_DIGESTS// }" ]]; then
+              echo "Stable release is missing prevalidated Windows installer digests." >&2
+              return 1
+            fi
+
+            windows_node_run_id="$(dispatch_workflow windows-node-release.yml \
+              -f tag="${RELEASE_TAG}" \
+              -f windows_node_tag="${WINDOWS_NODE_TAG}" \
+              -f expected_installer_digests="${WINDOWS_NODE_INSTALLER_DIGESTS}")"
+            echo "- Windows Node release run ID: \`${windows_node_run_id}\`" >> "$GITHUB_STEP_SUMMARY"
+            wait_for_run windows-node-release.yml "${windows_node_run_id}"
+          }
+
           upload_dependency_evidence_release_asset() {
             local release_version download_dir asset_path asset_name artifact_name
             release_version="${RELEASE_TAG#v}"
@@ -913,13 +1112,14 @@ jobs:
           }
 
           append_release_proof_to_github_release() {
-            local release_version body_file notes_file tarball integrity telegram_line clawhub_line clawhub_bootstrap_line clawhub_runtime_state_path
+            local release_version body_file notes_file evidence_path tarball integrity telegram_line clawhub_line clawhub_bootstrap_line clawhub_runtime_state_path windows_line
 
             release_version="${RELEASE_TAG#v}"
             body_file="${RUNNER_TEMP}/release-body.md"
             notes_file="${RUNNER_TEMP}/release-notes-with-proof.md"
-            tarball="$(npm view "openclaw@${release_version}" dist.tarball --json | jq -r '.')"
-            integrity="$(npm view "openclaw@${release_version}" dist.integrity --json | jq -r '.')"
+            evidence_path="${POSTPUBLISH_EVIDENCE_DIR}/release-postpublish-evidence.json"
+            tarball="$(jq -er '.openclawNpmTarball | select(type == "string" and length > 0)' "${evidence_path}")"
+            integrity="$(jq -er '.openclawNpmIntegrity | select(type == "string" and length > 0)' "${evidence_path}")"
             gh release view "${RELEASE_TAG}" --repo "$GITHUB_REPOSITORY" --json body --jq .body > "${body_file}"
 
             if [[ -n "${NPM_TELEGRAM_RUN_ID// }" ]]; then
@@ -931,6 +1131,10 @@ jobs:
             write_clawhub_runtime_state false "${clawhub_runtime_state_path}"
             clawhub_line="$(jq -r '.proofLines.normal' "${clawhub_runtime_state_path}")"
             clawhub_bootstrap_line="$(jq -r '.proofLines.bootstrap' "${clawhub_runtime_state_path}")"
+            windows_line=""
+            if [[ -n "${windows_node_run_id// }" ]]; then
+              windows_line="- Windows Hub promotion: https://github.com/${GITHUB_REPOSITORY}/actions/runs/${windows_node_run_id} from openclaw/openclaw-windows-node@${WINDOWS_NODE_TAG}"
+            fi
 
             RELEASE_BODY_FILE="${body_file}" \
               RELEASE_NOTES_FILE="${notes_file}" \
@@ -948,6 +1152,7 @@ jobs:
               CLAWHUB_LINE="${clawhub_line}" \
               CLAWHUB_BOOTSTRAP_LINE="${clawhub_bootstrap_line}" \
               TELEGRAM_LINE="${telegram_line}" \
+              WINDOWS_LINE="${windows_line}" \
               node --input-type=module <<'NODE'
           import { readFileSync, writeFileSync } from "node:fs";
 
@@ -974,6 +1179,7 @@ jobs:
             process.env.CLAWHUB_BOOTSTRAP_LINE,
             `- OpenClaw npm publish: https://github.com/${process.env.RELEASE_REPO}/actions/runs/${process.env.OPENCLAW_NPM_RUN_ID}`,
             process.env.TELEGRAM_LINE,
+            ...(process.env.WINDOWS_LINE ? [process.env.WINDOWS_LINE] : []),
           ].join("\n");
 
           const withoutOldProof = body.replace(/\n?### Release verification\n[\s\S]*?(?=\n### |\n## |$)/, "");
@@ -998,6 +1204,9 @@ jobs:
             else
               echo "- OpenClaw npm publish: skipped by input"
             fi
+            if is_stable_release && [[ "${PUBLISH_OPENCLAW_NPM}" == "true" ]]; then
+              echo "- Windows Hub promotion: required before the GitHub release can be published"
+            fi
             if [[ "${WAIT_FOR_CLAWHUB}" == "true" ]]; then
               echo "- Workflow completion waits for ClawHub"
             else
@@ -1142,6 +1351,7 @@ jobs:
 
           failed=0
           openclaw_failed=0
+          windows_node_run_id=""
           if [[ -n "${openclaw_pid}" ]] && ! wait "${openclaw_pid}"; then
             failed=1
             openclaw_failed=1
@@ -1172,6 +1382,9 @@ jobs:
             fi
             create_or_update_github_release
             upload_dependency_evidence_release_asset
+            if ! promote_windows_release_assets; then
+              failed=1
+            fi
             append_release_proof_to_github_release
             if [[ "${failed}" == "0" ]]; then
               publish_github_release

.github/workflows/qa-live-transports-convex.yml

@@ -532,7 +532,6 @@ jobs:
           OPENCLAW_QA_CONVEX_SECRET_CI: ${{ secrets.OPENCLAW_QA_CONVEX_SECRET_CI }}
           OPENCLAW_QA_CREDENTIAL_ACQUIRE_TIMEOUT_MS: "1800000"
           OPENCLAW_QA_REDACT_PUBLIC_METADATA: "1"
-          OPENCLAW_QA_TELEGRAM_CAPTURE_CONTENT: "1"
           INPUT_SCENARIO: ${{ github.event_name == 'workflow_dispatch' && inputs.scenario || '' }}
         run: |
           set -euo pipefail

.github/workflows/stale.yml

@@ -68,7 +68,7 @@ jobs:
           days-before-pr-close: 7
           stale-issue-label: stale
           stale-pr-label: stale
-          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle
+          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle,clawsweeper:queueable-fix,clawsweeper:source-repro,clawsweeper:fix-shape-clear
           exempt-pr-labels: maintainer,no-stale,bad-barnacle
           operations-per-run: 2000
           ascending: true
@@ -100,7 +100,7 @@ jobs:
           days-before-pr-stale: -1
           days-before-pr-close: -1
           stale-issue-label: stale
-          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle
+          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle,clawsweeper:queueable-fix,clawsweeper:source-repro,clawsweeper:fix-shape-clear
           operations-per-run: 2000
           ascending: true
           include-only-assigned: true
@@ -172,7 +172,7 @@ jobs:
           days-before-pr-close: 7
           stale-issue-label: stale
           stale-pr-label: stale
-          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle
+          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle,clawsweeper:queueable-fix,clawsweeper:source-repro,clawsweeper:fix-shape-clear
           exempt-pr-labels: maintainer,no-stale,bad-barnacle
           operations-per-run: 2000
           ascending: true
@@ -203,7 +203,7 @@ jobs:
           days-before-pr-stale: -1
           days-before-pr-close: -1
           stale-issue-label: stale
-          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle
+          exempt-issue-labels: enhancement,maintainer,pinned,security,no-stale,bad-barnacle,clawsweeper:queueable-fix,clawsweeper:source-repro,clawsweeper:fix-shape-clear
           operations-per-run: 2000
           ascending: true
           include-only-assigned: true
@@ -277,6 +277,9 @@ jobs:
               "security",
               "no-stale",
               "bad-barnacle",
+              "clawsweeper:queueable-fix",
+              "clawsweeper:source-repro",
+              "clawsweeper:fix-shape-clear",
             ]);
             const prExemptLabels = new Set(["maintainer", "no-stale", "bad-barnacle"]);
             const maintainerAssociations = new Set(["OWNER", "MEMBER", "COLLABORATOR"]);

.github/workflows/windows-node-release.yml

@@ -8,9 +8,12 @@ on:
         required: true
         type: string
       windows_node_tag:
-        description: openclaw-windows-node release tag to promote, or latest
+        description: Exact openclaw-windows-node release tag to promote, for example v0.6.3
+        required: true
+        type: string
+      expected_installer_digests:
+        description: Compact JSON map of installer asset names to pinned source sha256 digests
         required: true
-        default: latest
         type: string
 
 permissions:
@@ -31,46 +34,129 @@ jobs:
         env:
           RELEASE_TAG: ${{ inputs.tag }}
           WINDOWS_NODE_TAG: ${{ inputs.windows_node_tag }}
+          EXPECTED_INSTALLER_DIGESTS: ${{ inputs.expected_installer_digests }}
           GH_TOKEN: ${{ github.token }}
         run: |
           if ($env:RELEASE_TAG -notmatch '^v[0-9]{4}\.[1-9][0-9]*\.[1-9][0-9]*((-(alpha|beta)\.[1-9][0-9]*)|(-[1-9][0-9]*))?$') {
             throw "Invalid OpenClaw release tag: $env:RELEASE_TAG"
           }
-          if ($env:WINDOWS_NODE_TAG -ne "latest" -and $env:WINDOWS_NODE_TAG -notmatch '^v[0-9]+\.[0-9]+\.[0-9]+([-.][0-9A-Za-z.-]+)?$') {
-            throw "Invalid openclaw-windows-node release tag: $env:WINDOWS_NODE_TAG"
+          $stableRelease = -not (
+            $env:RELEASE_TAG.Contains("-alpha.") -or
+            $env:RELEASE_TAG.Contains("-beta.")
+          )
+          if ($env:WINDOWS_NODE_TAG -notmatch '^v[0-9]+\.[0-9]+\.[0-9]+([-.][0-9A-Za-z]+([.-][0-9A-Za-z]+)*)?$') {
+            throw "windows_node_tag must be an explicit openclaw-windows-node release tag, not latest: $env:WINDOWS_NODE_TAG"
+          }
+
+          try {
+            $expectedDigests = $env:EXPECTED_INSTALLER_DIGESTS | ConvertFrom-Json -AsHashtable
+          } catch {
+            throw "expected_installer_digests must be a JSON object: $_"
+          }
+          # Add future signed installer names, such as MSIX x64/ARM64, here.
+          $requiredInstallerNames = @(
+            "OpenClawCompanion-Setup-x64.exe",
+            "OpenClawCompanion-Setup-arm64.exe"
+          )
+          $allowedTargetCompanionAssetNames = @(
+            $requiredInstallerNames
+            "OpenClawCompanion-SHA256SUMS.txt"
+          )
+          if ($expectedDigests.Count -ne $requiredInstallerNames.Count) {
+            throw "expected_installer_digests must contain exactly the current installer asset contract."
+          }
+          foreach ($name in $requiredInstallerNames) {
+            $digest = [string]$expectedDigests[$name]
+            if ($digest -notmatch '^sha256:[A-Fa-f0-9]{64}$') {
+              throw "expected_installer_digests is missing a valid pinned digest for $name."
+            }
+          }
+
+          $targetRelease = gh release view $env:RELEASE_TAG --repo $env:GITHUB_REPOSITORY --json tagName,isDraft,isPrerelease,assets,url | ConvertFrom-Json
+          if ($targetRelease.tagName -ne $env:RELEASE_TAG) {
+            throw "OpenClaw release tag mismatch: expected $env:RELEASE_TAG, got $($targetRelease.tagName)"
+          }
+          $unexpectedTargetCompanionAssets = @(
+            $targetRelease.assets |
+              Where-Object {
+                $_.name.StartsWith("OpenClawCompanion-") -and
+                $_.name -notin $allowedTargetCompanionAssetNames
+              } |
+              ForEach-Object name |
+              Sort-Object
+          )
+          if ($unexpectedTargetCompanionAssets.Count -ne 0) {
+            throw "Target OpenClaw release contains unexpected OpenClawCompanion assets before upload: $($unexpectedTargetCompanionAssets -join ', ')"
+          }
+
+          $sourceRelease = gh release view $env:WINDOWS_NODE_TAG --repo openclaw/openclaw-windows-node --json tagName,isDraft,isPrerelease,assets,url | ConvertFrom-Json
+          if ($sourceRelease.tagName -ne $env:WINDOWS_NODE_TAG) {
+            throw "Windows source release tag mismatch: expected $env:WINDOWS_NODE_TAG, got $($sourceRelease.tagName)"
+          }
+          if ($sourceRelease.isDraft) {
+            throw "Windows source release must be published: $($sourceRelease.url)"
+          }
+          if ($stableRelease -and $sourceRelease.isPrerelease) {
+            throw "Stable OpenClaw releases require a non-prerelease Windows source release: $($sourceRelease.url)"
+          }
+          foreach ($name in $requiredInstallerNames) {
+            $sourceAssets = @($sourceRelease.assets | Where-Object name -eq $name)
+            if ($sourceAssets.Count -ne 1) {
+              throw "Windows source release must contain exactly one required asset $name; found $($sourceAssets.Count)."
+            }
+            if ([string]$sourceAssets[0].digest -ne [string]$expectedDigests[$name]) {
+              throw "Windows source release asset digest does not match the pinned digest: $name"
+            }
           }
-          gh release view $env:RELEASE_TAG --repo $env:GITHUB_REPOSITORY | Out-Null
 
       - name: Download Windows Hub release installers
         shell: pwsh
         env:
           WINDOWS_NODE_TAG: ${{ inputs.windows_node_tag }}
+          EXPECTED_INSTALLER_DIGESTS: ${{ inputs.expected_installer_digests }}
           GH_TOKEN: ${{ github.token }}
         run: |
           New-Item -ItemType Directory -Force -Path dist | Out-Null
-          $tagArgs = @()
-          if ($env:WINDOWS_NODE_TAG -ne "latest") {
-            $tagArgs += $env:WINDOWS_NODE_TAG
-          }
-          gh release download @tagArgs `
-            --repo openclaw/openclaw-windows-node `
-            --pattern "OpenClawCompanion-Setup-*.exe" `
-            --dir dist
-
-          $expected = @(
-            "dist/OpenClawCompanion-Setup-x64.exe",
-            "dist/OpenClawCompanion-Setup-arm64.exe"
+          # Add future signed installer patterns, such as MSIX x64/ARM64, here.
+          # Every matched installer is signature-checked, checksummed, and promoted.
+          $installerPatterns = @(
+            "OpenClawCompanion-Setup-x64.exe",
+            "OpenClawCompanion-Setup-arm64.exe"
           )
-          foreach ($file in $expected) {
-            if (-not (Test-Path -LiteralPath $file)) {
-              throw "Missing expected Windows installer: $file"
+          $downloadArgs = @(
+            $env:WINDOWS_NODE_TAG,
+            "--repo", "openclaw/openclaw-windows-node",
+            "--dir", "dist"
+          )
+          foreach ($pattern in $installerPatterns) {
+            $downloadArgs += @("--pattern", $pattern)
+          }
+          gh release download @downloadArgs
+          if ($LASTEXITCODE -ne 0) {
+            throw "Failed to download Windows release assets from $env:WINDOWS_NODE_TAG."
+          }
+
+          foreach ($pattern in $installerPatterns) {
+            $patternMatches = @(Get-ChildItem -LiteralPath dist -File | Where-Object Name -Like $pattern)
+            if ($patternMatches.Count -ne 1) {
+              throw "Expected exactly one Windows installer matching '$pattern', found $($patternMatches.Count)."
+            }
+          }
+
+          $expectedDigests = $env:EXPECTED_INSTALLER_DIGESTS | ConvertFrom-Json -AsHashtable
+          foreach ($file in Get-ChildItem -LiteralPath dist -File) {
+            $expectedHash = ([string]$expectedDigests[$file.Name]) -replace '^sha256:', ''
+            $actualHash = (Get-FileHash -Algorithm SHA256 -LiteralPath $file.FullName).Hash
+            if ($actualHash -ne $expectedHash) {
+              throw "Downloaded Windows source asset does not match pinned digest: $($file.Name)"
             }
           }
 
       - name: Verify Authenticode signatures
         shell: pwsh
         run: |
-          Get-ChildItem -LiteralPath dist -Filter "OpenClawCompanion-Setup-*.exe" | ForEach-Object {
+          $expectedSignerSubject = "CN=OpenClaw Foundation, O=OpenClaw Foundation, L=Mill Valley, S=California, C=US"
+          Get-ChildItem -LiteralPath dist -File | ForEach-Object {
             $signature = Get-AuthenticodeSignature -LiteralPath $_.FullName
             if ($signature.Status -ne "Valid") {
               throw "$($_.Name) Authenticode signature was $($signature.Status)."
@@ -78,6 +164,9 @@ jobs:
             if (-not $signature.SignerCertificate) {
               throw "$($_.Name) has no signer certificate."
             }
+            if ($signature.SignerCertificate.Subject -ne $expectedSignerSubject) {
+              throw "$($_.Name) has unexpected signer subject $($signature.SignerCertificate.Subject)."
+            }
             [pscustomobject]@{
               File = $_.Name
               Signer = $signature.SignerCertificate.Subject
@@ -88,7 +177,7 @@ jobs:
       - name: Write SHA-256 manifest
         shell: pwsh
         run: |
-          Get-ChildItem -LiteralPath dist -Filter "OpenClawCompanion-Setup-*.exe" |
+          Get-ChildItem -LiteralPath dist -File |
             Sort-Object Name |
             ForEach-Object {
               $hash = Get-FileHash -Algorithm SHA256 -LiteralPath $_.FullName
@@ -101,12 +190,81 @@ jobs:
           RELEASE_TAG: ${{ inputs.tag }}
           GH_TOKEN: ${{ github.token }}
         run: |
-          gh release upload $env:RELEASE_TAG `
-            dist/OpenClawCompanion-Setup-x64.exe `
-            dist/OpenClawCompanion-Setup-arm64.exe `
-            dist/OpenClawCompanion-SHA256SUMS.txt `
-            --repo $env:GITHUB_REPOSITORY `
-            --clobber
+          $releaseAssets = @(Get-ChildItem -LiteralPath dist -File | Sort-Object Name | ForEach-Object FullName)
+          gh release upload $env:RELEASE_TAG @releaseAssets --repo $env:GITHUB_REPOSITORY --clobber
+          if ($LASTEXITCODE -ne 0) {
+            throw "Failed to upload Windows release assets to $env:RELEASE_TAG."
+          }
+
+      - name: Verify promoted release asset contract
+        shell: pwsh
+        env:
+          RELEASE_TAG: ${{ inputs.tag }}
+          GH_TOKEN: ${{ github.token }}
+        run: |
+          New-Item -ItemType Directory -Force -Path verified | Out-Null
+          $expectedAssets = @(Get-ChildItem -LiteralPath dist -File | Sort-Object Name)
+          $expectedCompanionAssetNames = @($expectedAssets | ForEach-Object Name | Sort-Object)
+          $targetRelease = gh release view $env:RELEASE_TAG --repo $env:GITHUB_REPOSITORY --json assets | ConvertFrom-Json
+          $actualCompanionAssetNames = @(
+            $targetRelease.assets |
+              Where-Object { $_.name.StartsWith("OpenClawCompanion-") } |
+              ForEach-Object name |
+              Sort-Object
+          )
+          $assetContractDiff = @(
+            Compare-Object `
+              -ReferenceObject $expectedCompanionAssetNames `
+              -DifferenceObject $actualCompanionAssetNames
+          )
+          if (
+            $actualCompanionAssetNames.Count -ne $expectedCompanionAssetNames.Count -or
+            $assetContractDiff.Count -ne 0
+          ) {
+            throw "Promoted OpenClawCompanion asset names do not exactly match the current contract."
+          }
+
+          foreach ($asset in $expectedAssets) {
+            gh release download $env:RELEASE_TAG `
+              --repo $env:GITHUB_REPOSITORY `
+              --pattern $asset.Name `
+              --dir verified
+            if ($LASTEXITCODE -ne 0) {
+              throw "Failed to download promoted Windows release asset $($asset.Name)."
+            }
+          }
+
+          $manifestPath = "verified/OpenClawCompanion-SHA256SUMS.txt"
+          $manifestEntries = @(Get-Content -LiteralPath $manifestPath | ForEach-Object {
+            if ($_ -notmatch '^([A-Fa-f0-9]{64})  ([^\\/]+)$') {
+              throw "Invalid Windows SHA-256 manifest entry: $_"
+            }
+            [PSCustomObject]@{
+              Hash = $Matches[1]
+              Name = $Matches[2]
+            }
+          })
+          $expectedInstallerNames = @(
+            $expectedAssets |
+              Where-Object Name -ne "OpenClawCompanion-SHA256SUMS.txt" |
+              ForEach-Object Name
+          )
+          $manifestInstallerNames = @($manifestEntries | ForEach-Object Name | Sort-Object)
+          $contractDiff = @(
+            Compare-Object `
+              -ReferenceObject $expectedInstallerNames `
+              -DifferenceObject $manifestInstallerNames
+          )
+          if ($contractDiff.Count -ne 0) {
+            throw "Promoted Windows SHA-256 manifest does not match the installer asset contract."
+          }
+
+          foreach ($entry in $manifestEntries) {
+            $hash = (Get-FileHash -Algorithm SHA256 -LiteralPath "verified/$($entry.Name)").Hash
+            if ($hash -ne $entry.Hash) {
+              throw "Promoted Windows release asset checksum mismatch: $($entry.Name)"
+            }
+          }
 
       - name: Summary
         shell: pwsh
@@ -119,8 +277,9 @@ jobs:
 
           OpenClaw release: $env:RELEASE_TAG
           Source release: openclaw/openclaw-windows-node@$env:WINDOWS_NODE_TAG
-
-          - https://github.com/openclaw/openclaw/releases/download/$env:RELEASE_TAG/OpenClawCompanion-Setup-x64.exe
-          - https://github.com/openclaw/openclaw/releases/download/$env:RELEASE_TAG/OpenClawCompanion-Setup-arm64.exe
-          - https://github.com/openclaw/openclaw/releases/download/$env:RELEASE_TAG/OpenClawCompanion-SHA256SUMS.txt
           "@ >> $env:GITHUB_STEP_SUMMARY
+          Get-ChildItem -LiteralPath dist -File |
+            Sort-Object Name |
+            ForEach-Object {
+              "- https://github.com/openclaw/openclaw/releases/download/$env:RELEASE_TAG/$($_.Name)"
+            } >> $env:GITHUB_STEP_SUMMARY

CHANGELOG.md

@@ -2,6 +2,38 @@
 
 Docs: https://docs.openclaw.ai
 
+## 2026.6.8
+
+### Highlights
+
+- Telegram and WhatsApp channel delivery are richer and less brittle: Telegram can send structured rich text with tables, lists, expandable blockquotes, preserved intentional line breaks, prompt-preserving CLI backend delivery, retired native draft migration, and safer rich-media boundaries, while WhatsApp now honors configured ACP bindings. (#92679, #93164, #84082, #89421, #92513) Thanks @obviyus, @jzakirov, @spacegeologist, and @TurboTheTurtle.
+- Agent and Gateway recovery is sharper across account-scoped DM sends, generated media completions, auto-reply message-tool final replies, reset archive fallback reads, restart shutdown aborts, yielded subagent pauses, trusted subagent thinking override fallback, yielded cron media, heartbeat dedupe, session identity prompts, and unknown OpenAI agent selector rejection. (#92788, #91246, #92879, #91357, #92631, #92412, #92146, #91287, #92468, #92510) Thanks @yetval, @TurboTheTurtle, @masatohoshino, @CadanHu, @ooiuuii, @openperf, @IWhatsskill, @ZengWen-DT, and @zhangguiping-xydt.
+- Provider/model handling expands and tightens with GLM-5.2, Claude Haiku 4.5 catalog rows, OpenRouter and Google Vertex provider-prefix normalization, managed SecretRef auth, OAuth image-default routing through Codex, bounded model browse discovery, LM Studio binary thinking-off delivery, storeless OpenAI Responses replay gating, invalid OpenAI reasoning-signature and genericized Anthropic thinking-signature recovery, Claude 4.5 Copilot tool-streaming safety, and OpenAI/Anthropic-family payload quarantine for unreadable or post-hook tool schemas. (#92796, #90116, #92627, #91218, #90686, #92824, #92247, #92002, #90706, #92941, #92201, #92916, #75393, #92908, #92921, #92928) Thanks @arkyu2077, @liuhao1024, @bymle, @rohitjavvadi, @nxmxbbd, @bek91, @samson910022, @mmyzwl, @CarlCapital, @snowzlm, @Kailigithub, and @vincentkoc.
+- `/usage` and reply payload hooks now have a native full footer renderer, default template, fixed-decimal formatting, credential-aware limits, better partial-count handling, and warnings for broken templates instead of silent bad output. (#92657, #89835, #89629) Thanks @Marvinthebored.
+- UI and mobile flows are steadier: workspace files can collapse and start collapsed, WebChat backscroll survives streaming, the sidebar session picker remains interactive above the desktop workbench, reset soft args survive UI dispatch, stale dashboard session parent lineage is preserved, and iOS reconnects stale foreground gateways. (#92779, #92622, #92705, #91353, #90658, #92552) Thanks @shakkernerd, @TurboTheTurtle, @NianJiuZst, @zhouhe-xydt, @luoyanglang, and @Solvely-Colin.
+- Memory, state, and diagnostics recover cleaner: oversized OpenAI embedding batches split before 431s, QMD memory search stays available in transient mode, SQLite avoids WAL on NFS state volumes, stuck-session recovery scheduling no longer resets warning backoff, full memory reindexes preserve rollback/cache recovery, raw Memory Wiki source pages stop looking malformed, and Infinity chunk limits stay genuinely unbounded. (#92650, #92618, #92639, #91247, #92752, #92881, #59137, #92876, #69700, #92735) Thanks @mushuiyu886, @TurboTheTurtle, @849261680, @gnanam1990, @TSHOGX, @arlen8411, and @yhterrance.
+
+### Changes
+
+- Providers/models: add GLM-5.2 support and Claude Haiku 4.5 catalog entries while keeping provider-qualified model IDs normalized across OpenRouter and Google Vertex paths. (#92796, #90116, #92627, #91218) Thanks @arkyu2077, @liuhao1024, and @bymle.
+- Web search: keep key-free providers such as Parallel Free, DuckDuckGo, Ollama, and Codex Hosted Search as explicit opt-ins instead of selecting them automatically when no API-backed provider is configured. (#93616) Thanks @davemorin and @vincentkoc.
+- Channel plugins: ship Telegram rich-message delivery and WhatsApp ACP binding support, including preserved intentional line breaks, rich prompt handoff to CLI backends, and transport fixtures for richer drafts. (#92679, #93164, #92513) Thanks @obviyus and @TurboTheTurtle.
+- Agent commands: support `/btw` in CLI-backed sessions and keep CLI usage-error exits classified as usage failures instead of successful runs. (#92669, #92162) Thanks @joshavant and @Pandah97.
+- Usage hooks: add built-in full footer rendering, default footer templates, per-turn usage state, credential-aware limits, and fixed-decimal formatting for usage-bar templates. (#92657, #89835, #89629) Thanks @Marvinthebored.
+- Docs and operator guidance: document node config examples, clarify before-install hook scope, correct agent default concurrency comments, refresh ZAI provider docs, and update channel/group docs for current Telegram and WhatsApp behavior. (#92677, #92766, #92695) Thanks @liuhao1024, @sallyom, and @ArielSmoliar.
+
+### Fixes
+
+- Channels and delivery: preserve account-scoped DM channel send policy, intentional rich-message line breaks in Telegram and status output, rich Telegram final replies, rich Telegram tables and lists, Telegram thread-create CLI remapping, Feishu dynamic-agent routes after persisted binding reuse, Slack outbound `message_sent` hooks, contributed message-tool schema optionality, same-channel generated media completions, and channel chunking around surrogate pairs and Infinity limits. (#92788, #93164, #92679, #89421, #89943, #42837, #92814, #91137, #91246, #92735) Thanks @yetval, @obviyus, @spacegeologist, @rishitamrakar, @liuhao1024, @lundog, @TurboTheTurtle, and @yhterrance.
+- Discord: give generated auto-thread titles a 60-second timeout and 4,096-token reasoning-model output budget, clamped to the selected model output cap. (#64734) Thanks @hanamizuki.
+- Agent, cron, and Gateway runtime: mark active main sessions before restart shutdown aborts, pause yielded subagent runs whose terminal also signals abort, clamp trusted subagent thinking overrides through provider/model fallback, preserve yielded media completions, deliver channel message-tool final replies through auto-reply while hiding internal delivery hints, restore reset archive fallback reads when active async transcripts are missing, de-duplicate main-session heartbeat events, expose session identity in runtime prompts, reject unknown OpenAI agent selectors, keep generated media completions, slash-command block replies, and trajectory export commands in WebChat, and require admin privileges for HTTP session/model override surfaces. (#91357, #92631, #92412, #92146, #92879, #91287, #92468, #92510, #91246, #92651, #92646) Thanks @ooiuuii, @openperf, @IWhatsskill, @masatohoshino, @CadanHu, @ZengWen-DT, @zhangguiping-xydt, and @TurboTheTurtle.
+- Providers and model replay: preserve storeless OpenAI Responses replay compatibility, recover invalid OpenAI reasoning signatures and genericized Anthropic thinking-signature replay errors, route OAuth image defaults through Codex for eligible OpenAI profiles, avoid eager tool streaming for Claude 4.5 in Copilot, quarantine unreadable and post-hook OpenAI/Anthropic-family tool schemas without broadening allowed tool choices, deliver explicit thinking-off requests to LM Studio binary-thinking models, honor profile auth for SecretRef model entries, bound model browsing, strip provider prefixes where runtimes need bare IDs, and surface nested embedding fetch failures. (#90706, #92941, #92201, #92916, #92824, #75393, #92908, #92921, #92928, #92002, #90686, #92247, #92627, #91218, #92628) Thanks @snowzlm, @mmyzwl, @CarlCapital, @bek91, @Kailigithub, @vincentkoc, @rohitjavvadi, @samson910022, @nxmxbbd, @liuhao1024, @bymle, and @mushuiyu886.
+- Memory, state, diagnostics, and config: split header-too-large embedding batches, keep QMD memory search enabled in transient mode, avoid SQLite WAL on NFS volumes, preserve recovery scheduling outside stuck-session warning backoff, preserve full-reindex rollback/cache recovery, treat raw Memory Wiki source pages as source evidence, and keep shell environment fallbacks contained in config write tests. (#92650, #92618, #92639, #91247, #92752, #92881, #59137, #92876, #69700) Thanks @mushuiyu886, @TurboTheTurtle, @849261680, @gnanam1990, @TSHOGX, and @arlen8411.
+- UI/mobile/TUI: preserve dashboard session parent lineage, WebChat backscroll, reset soft command args, sidebar session picker interactivity, collapsed workspace files, resolved `/model` confirmation refs, stale foreground iOS Gateway reconnects, and paused setup-parent stdin after inherited-stdio child exit. (#90658, #92622, #91353, #92705, #92779, #92773, #92552, #93159) Thanks @luoyanglang, @TurboTheTurtle, @zhouhe-xydt, @NianJiuZst, @shakkernerd, @NarahariRaghava, @Solvely-Colin, and @fuller-stack-dev.
+- Plugins and updates: repair missing required platform packages during managed plugin installs and updates, including omitted Codex platform binaries.
+- Dependencies: update Hono to 4.12.25 so published OpenClaw and ACPX packages use the patched runtime.
+- Release and test reliability: extend slow Gateway/full-suite watchdogs, split local full-suite shards when throttled, stabilize plugin auth marker fixtures, avoid brittle provider-ref error text, fold Telegram RTT sampling into live QA evidence, simplify QA scorecard mappings around canonical coverage IDs, keep QA Lab bootstrap selection assertions aligned with flow-only scenarios, skip QA coverage artifact consumers when runtime parity producer status is not green, keep Feishu lifecycle release checks pointed at the active fixture config, isolate trajectory-export live seed turns from Codex-native shell approvals, preserve release-check child refs while pinning expected SHAs, widen live OpenAI TTS budgets for slower provider responses, and avoid false downgrade prompts for unresolved latest-tag updates. (#92652, #92550, #92558, #92911) Thanks @RomneyDa and @Andy312432.
+
 ## 2026.6.6
 
 ### Highlights

Dockerfile

@@ -138,7 +138,7 @@ ARG OPENCLAW_BUNDLED_PLUGIN_DIR
 # BuildKit cache mounts are not part of cached layers; seed tarballs for the
 # installed prod graph in the same step that runs offline prune.
 RUN --mount=type=cache,id=openclaw-pnpm-store,target=/root/.local/share/pnpm/store,sharing=locked \
-    pnpm list --prod --depth Infinity --json | node scripts/list-prod-store-packages.mjs | xargs -r pnpm store add && \
+    node scripts/list-prod-store-packages.mjs | xargs -r pnpm store add && \
     CI=true pnpm prune --prod \
       --config.offline=true \
       --config.supportedArchitectures.os=linux \
@@ -147,6 +147,10 @@ RUN --mount=type=cache,id=openclaw-pnpm-store,target=/root/.local/share/pnpm/sto
     OPENCLAW_EXTENSIONS="$OPENCLAW_EXTENSIONS" OPENCLAW_BUNDLED_PLUGIN_DIR="$OPENCLAW_BUNDLED_PLUGIN_DIR" node scripts/prune-docker-plugin-dist.mjs && \
     node scripts/postinstall-bundled-plugins.mjs && \
     find dist -type f \( -name '*.d.ts' -o -name '*.d.mts' -o -name '*.d.cts' -o -name '*.map' \) -delete && \
+    rm -rf \
+      /app/node_modules/openclaw \
+      /app/node_modules/.bin/openclaw \
+      /app/node_modules/.pnpm/openclaw@*/node_modules/openclaw && \
     node scripts/check-package-dist-imports.mjs /app
 
 # ── Runtime base image ──────────────────────────────────────────

apps/android/app/build.gradle.kts

@@ -65,8 +65,8 @@ android {
     applicationId = "ai.openclaw.app"
     minSdk = 31
     targetSdk = 36
-    versionCode = 2026060201
-    versionName = "2026.6.2"
+    versionCode = 2026060801
+    versionName = "2026.6.8"
     ndk {
       // Support all major ABIs — native libs are tiny (~47 KB per ABI)
       abiFilters += listOf("armeabi-v7a", "arm64-v8a", "x86", "x86_64")

apps/ios/CHANGELOG.md

@@ -1,5 +1,9 @@
 # OpenClaw iOS Changelog
 
+## 2026.6.8 - 2026-06-14
+
+Maintenance update for the current OpenClaw beta release.
+
 ## 2026.6.2 - 2026-06-02
 
 OpenClaw is now available on iPhone.

apps/ios/Config/Version.xcconfig

@@ -2,8 +2,8 @@
 // Source of truth: apps/ios/version.json
 // Generated by scripts/ios-sync-versioning.ts.
 
-OPENCLAW_IOS_VERSION = 2026.6.2
-OPENCLAW_MARKETING_VERSION = 2026.6.2
+OPENCLAW_IOS_VERSION = 2026.6.8
+OPENCLAW_MARKETING_VERSION = 2026.6.8
 OPENCLAW_BUILD_VERSION = 1
 
 #include? "../build/Version.xcconfig"

apps/ios/Sources/Model/NodeAppModel.swift

@@ -188,6 +188,7 @@ final class NodeAppModel {
     @ObservationIgnored private var backgroundGraceTaskTimer: Task<Void, Never>?
     private var backgroundReconnectSuppressed = false
     private var backgroundReconnectLeaseUntil: Date?
+    @ObservationIgnored private var foregroundGatewayResumeCheckInFlight = false
     private var lastSignificantLocationWakeAt: Date?
     @ObservationIgnored private let watchReplyCoordinator = WatchReplyCoordinator()
     private var watchExecApprovalPromptsByID: [String: ExecApprovalPrompt] = [:]
@@ -214,6 +215,7 @@ final class NodeAppModel {
     private static let watchExecApprovalBridgeStateKey = "watch.execApproval.bridge.state.v1"
     private static let backgroundAliveLastSuccessAtMsKey = "gateway.backgroundAlive.lastSuccessAtMs"
     private static let backgroundAliveLastTriggerKey = "gateway.backgroundAlive.lastTrigger"
+    private static let foregroundResumeHealthTimeoutSeconds = 1
 
     var cameraHUDText: String?
     var cameraHUDKind: CameraHUDKind?
@@ -417,9 +419,7 @@ final class NodeAppModel {
             self.isBackgrounded = false
             self.endBackgroundConnectionGracePeriod(reason: "scene_foreground")
             self.clearBackgroundReconnectSuppression(reason: "scene_foreground")
-            if self.operatorConnected {
-                self.startGatewayHealthMonitor()
-            }
+            var shouldStartGatewayHealthMonitor = self.operatorConnected
             if phase == .active {
                 self.voiceWake.resumeAfterExternalAudioCapture(wasSuspended: self.backgroundVoiceWakeSuspended)
                 self.backgroundVoiceWakeSuspended = false
@@ -444,6 +444,8 @@ final class NodeAppModel {
                 // iOS may suspend network sockets in background without a clean close.
                 // On foreground, force a fresh handshake to avoid "connected but dead" states.
                 if backgroundedFor >= 3.0 {
+                    shouldStartGatewayHealthMonitor = false
+                    self.foregroundGatewayResumeCheckInFlight = true
                     Task { [weak self] in
                         guard let self else { return }
                         let operatorWasConnected = await MainActor.run { self.operatorConnected }
@@ -452,31 +454,26 @@ final class NodeAppModel {
                             let healthy = await (try? self.operatorGateway.request(
                                 method: "health",
                                 paramsJSON: nil,
-                                timeoutSeconds: 2)) != nil
+                                timeoutSeconds: Self.foregroundResumeHealthTimeoutSeconds)) != nil
                             if healthy {
-                                await MainActor.run { self.startGatewayHealthMonitor() }
+                                await MainActor.run {
+                                    self.foregroundGatewayResumeCheckInFlight = false
+                                    self.startGatewayHealthMonitor()
+                                }
                                 return
                             }
                         }
 
-                        await self.operatorGateway.disconnect()
-                        await self.nodeGateway.disconnect()
                         await MainActor.run {
-                            guard !self.isAppleReviewDemoModeEnabled else { return }
-                            self.setOperatorConnected(false)
-                            self.gatewayConnected = false
-                            // Foreground recovery must actively restart the saved gateway config.
-                            // Disconnecting stale sockets alone can leave us idle if the old
-                            // reconnect tasks were suppressed or otherwise got stuck in background.
-                            self.gatewayStatusText = "Reconnecting…"
-                            self.talkMode.updateGatewayConnected(false)
-                            if let cfg = self.activeGatewayConnectConfig {
-                                self.applyGatewayConnectConfig(cfg)
-                            }
+                            self.foregroundGatewayResumeCheckInFlight = false
                         }
+                        await self.restartGatewaySessionsAfterForegroundStaleConnection()
                     }
                 }
             }
+            if shouldStartGatewayHealthMonitor {
+                self.startGatewayHealthMonitor()
+            }
         @unknown default:
             self.isBackgrounded = false
             self.endBackgroundConnectionGracePeriod(reason: "scene_unknown")
@@ -786,6 +783,12 @@ final class NodeAppModel {
 
     func refreshGatewayOverviewIfConnected() async {
         guard await self.isOperatorConnected() else { return }
+        if self.foregroundGatewayResumeCheckInFlight {
+            GatewayDiagnostics.log("gateway overview refresh deferred reason=foreground_resume_check")
+            try? await Task.sleep(
+                nanoseconds: UInt64(Self.foregroundResumeHealthTimeoutSeconds) * 1_000_000_000)
+            guard await self.isOperatorConnected(), !self.foregroundGatewayResumeCheckInFlight else { return }
+        }
         await self.refreshBrandingFromGateway()
         await self.refreshAgentsFromGateway()
     }
@@ -1986,12 +1989,33 @@ extension NodeAppModel {
     }
 
     func resetGatewaySessionsForForcedReconnect() async {
-        self.nodeGatewayTask?.cancel()
+        let nodeGatewayTask = self.nodeGatewayTask
+        let operatorGatewayTask = self.operatorGatewayTask
+        nodeGatewayTask?.cancel()
         self.nodeGatewayTask = nil
-        self.operatorGatewayTask?.cancel()
+        operatorGatewayTask?.cancel()
         self.operatorGatewayTask = nil
         await self.operatorGateway.disconnect()
         await self.nodeGateway.disconnect()
+        // Foreground recovery reuses the same config immediately after reset.
+        // Wait for canceled loops so their shutdown cleanup cannot clobber the new reconnect state.
+        if let operatorGatewayTask {
+            await operatorGatewayTask.value
+        }
+        if let nodeGatewayTask {
+            await nodeGatewayTask.value
+        }
+    }
+
+    private func restartGatewaySessionsAfterForegroundStaleConnection() async {
+        await self.resetGatewaySessionsForForcedReconnect()
+        guard !self.isAppleReviewDemoModeEnabled else { return }
+        self.setOperatorConnected(false)
+        self.gatewayConnected = false
+        self.gatewayStatusText = "Reconnecting…"
+        self.talkMode.updateGatewayConnected(false)
+        guard let cfg = self.activeGatewayConnectConfig else { return }
+        self.applyGatewayConnectConfig(cfg, forceReconnect: true)
     }
 
     func disconnectGateway() {
@@ -4826,6 +4850,10 @@ extension NodeAppModel {
         (self.nodeGatewayTask != nil, self.operatorGatewayTask != nil)
     }
 
+    func _test_restartGatewaySessionsAfterForegroundStaleConnection() async {
+        await self.restartGatewaySessionsAfterForegroundStaleConnection()
+    }
+
     func _test_handleSuccessfulBootstrapGatewayOnboarding() async {
         await self.handleSuccessfulBootstrapGatewayOnboarding(
             url: URL(string: "wss://gateway.example")!,

apps/ios/Tests/GatewayConnectionControllerTests.swift

@@ -356,6 +356,20 @@ import UIKit
         #expect(!appModel._test_hasGatewayLoopTasks().operator)
     }
 
+    @Test @MainActor func foregroundStaleConnectionRestartReappliesActiveGatewayConfig() async {
+        let appModel = NodeAppModel()
+        defer { appModel.disconnectGateway() }
+
+        let config = Self.makeGatewayConnectConfig()
+        appModel.applyGatewayConnectConfig(config)
+        await appModel._test_restartGatewaySessionsAfterForegroundStaleConnection()
+
+        #expect(appModel.gatewayStatusText == "Reconnecting…")
+        #expect(appModel.activeGatewayConnectConfig?.hasSameConnectionInputs(as: config) == true)
+        #expect(appModel._test_hasGatewayLoopTasks().node)
+        #expect(appModel._test_hasGatewayLoopTasks().operator)
+    }
+
     @Test @MainActor func loadLastConnectionReadsSavedValues() {
         let prior = KeychainStore.loadString(service: "ai.openclaw.gateway", account: "lastConnection")
         defer {

apps/ios/fastlane/metadata/en-US/release_notes.txt

@@ -1,3 +1 @@
-OpenClaw is now available on iPhone.
-
-Connect to your OpenClaw Gateway to chat with your assistant, use realtime Talk mode, review approvals, share content from iOS, and bring device capabilities like camera, location, screen, and notifications into your private automation workflows.
+Maintenance update for the current OpenClaw beta release.

apps/ios/version.json

@@ -1,3 +1,3 @@
 {
-  "version": "2026.6.2"
+  "version": "2026.6.8"
 }

apps/macos/Package.resolved

@@ -1,5 +1,5 @@
 {
-  "originHash" : "035a4fe955164c62c1628de75f6437a14443a947eea2a1b0176ba484d6fde6f8",
+  "originHash" : "ae9f37f50cff0d32d189e60948f61e2fa1704e997a6ef4ad5e37f6a11c165ea4",
   "pins" : [
     {
       "identity" : "axorcist",
@@ -42,8 +42,8 @@
       "kind" : "remoteSourceControl",
       "location" : "https://github.com/steipete/Peekaboo.git",
       "state" : {
-        "revision" : "3a56ed2aa769bfefb5a78722dfce3c34088cfba1",
-        "version" : "3.4.0"
+        "revision" : "ee0e3185431788dad533ffca77cd75315aa3d26f",
+        "version" : "3.4.1"
       }
     },
     {
@@ -51,8 +51,8 @@
       "kind" : "remoteSourceControl",
       "location" : "https://github.com/sparkle-project/Sparkle",
       "state" : {
-        "revision" : "6276ba2b404829d139c45ff98427cf90e2efc59b",
-        "version" : "2.9.2"
+        "revision" : "d46d456107feacc80711b21847b82b07bd9fb46e",
+        "version" : "2.9.3"
       }
     },
     {
@@ -78,8 +78,8 @@
       "kind" : "remoteSourceControl",
       "location" : "https://github.com/apple/swift-log.git",
       "state" : {
-        "revision" : "2aed77ae5ec9a86d8fe42c12275e4c2653a286ee",
-        "version" : "1.13.1"
+        "revision" : "92448c359f00ebe36ae97d3bd9086f13c7692b5a",
+        "version" : "1.13.2"
       }
     },
     {

apps/macos/Package.swift

@@ -19,7 +19,7 @@ let package = Package(
         .package(url: "https://github.com/swiftlang/swift-subprocess.git", from: "0.4.0"),
         .package(url: "https://github.com/apple/swift-log.git", from: "1.10.1"),
         .package(url: "https://github.com/sparkle-project/Sparkle", from: "2.9.0"),
-        .package(url: "https://github.com/steipete/Peekaboo.git", exact: "3.4.0"),
+        .package(url: "https://github.com/steipete/Peekaboo.git", exact: "3.4.1"),
         .package(path: "../shared/OpenClawKit"),
         .package(path: "../swabble"),
     ],

apps/macos/Sources/OpenClaw/Resources/Info.plist

@@ -15,9 +15,9 @@
     <key>CFBundlePackageType</key>
     <string>APPL</string>
     <key>CFBundleShortVersionString</key>
-    <string>2026.6.2</string>
+    <string>2026.6.8</string>
     <key>CFBundleVersion</key>
-    <string>2026060200</string>
+    <string>2026060800</string>
     <key>CFBundleIconFile</key>
     <string>OpenClaw</string>
     <key>CFBundleURLTypes</key>

apps/macos/Sources/OpenClaw/VoiceWakeOverlayController+Window.swift

@@ -92,7 +92,13 @@ extension VoiceWakeOverlayController {
 
         let contentHeight = ceil(used.height + (textInset.height * 2))
         let total = contentHeight + self.verticalPadding * 2
-        self.model.isOverflowing = total > self.maxHeight
+        // Defer the overflow state mutation to break the SwiftUI onChange → measuredHeight →
+        // isOverflowing → re-render → onChange synchronous render loop (fixes #43480).
+        let overflowing = total > self.maxHeight
+        DispatchQueue.main.async { [weak self] in
+            guard let self, self.model.isOverflowing != overflowing else { return }
+            self.model.isOverflowing = overflowing
+        }
         return max(self.minHeight, min(total, self.maxHeight))
     }
 

apps/macos/Tests/OpenClawIPCTests/ExecApprovalsStoreRefactorTests.swift

@@ -4,14 +4,64 @@ import Testing
 
 @Suite(.serialized)
 struct ExecApprovalsStoreRefactorTests {
+    private var realTemporaryDirectory: URL {
+        let path = FileManager().temporaryDirectory.path
+        if path.hasPrefix("/var/") {
+            return URL(fileURLWithPath: "/private\(path)", isDirectory: true)
+        }
+        return FileManager().temporaryDirectory.resolvingSymlinksInPath()
+    }
+
+    private func withLockedEnv(
+        _ values: [String: String?],
+        _ body: () async throws -> Void) async throws
+    {
+        func restoreEnv(_ values: [String: String?]) {
+            for (key, value) in values {
+                if let value {
+                    setenv(key, value, 1)
+                } else {
+                    unsetenv(key)
+                }
+            }
+        }
+
+        await TestIsolationLock.shared.acquire()
+        var previousEnv: [String: String?] = [:]
+        for (key, value) in values {
+            previousEnv[key] = getenv(key).map { String(cString: $0) }
+            if let value {
+                setenv(key, value, 1)
+            } else {
+                unsetenv(key)
+            }
+        }
+
+        do {
+            try await body()
+            restoreEnv(previousEnv)
+            await TestIsolationLock.shared.release()
+        } catch {
+            restoreEnv(previousEnv)
+            await TestIsolationLock.shared.release()
+            throw error
+        }
+    }
+
     private func withTempStateDir(
         _ body: @escaping @Sendable (URL) async throws -> Void) async throws
     {
-        let stateDir = FileManager().temporaryDirectory
+        let root = self.realTemporaryDirectory
             .appendingPathComponent("openclaw-state-\(UUID().uuidString)", isDirectory: true)
-        defer { try? FileManager().removeItem(at: stateDir) }
+        let home = root.appendingPathComponent("home", isDirectory: true)
+        let stateDir = root.appendingPathComponent("state", isDirectory: true)
+        defer { try? FileManager().removeItem(at: root) }
+        try Self.seedCurrentApprovalsFile(in: stateDir)
 
-        try await TestIsolation.withEnvValues(["OPENCLAW_STATE_DIR": stateDir.path]) {
+        try await self.withLockedEnv([
+            "OPENCLAW_HOME": home.path,
+            "OPENCLAW_STATE_DIR": stateDir.path,
+        ]) {
             try await body(stateDir)
         }
     }
@@ -19,13 +69,13 @@ struct ExecApprovalsStoreRefactorTests {
     private func withTempHomeAndStateDir(
         _ body: @escaping @Sendable (URL, URL) async throws -> Void) async throws
     {
-        let root = FileManager().temporaryDirectory
+        let root = self.realTemporaryDirectory
             .appendingPathComponent("openclaw-home-state-\(UUID().uuidString)", isDirectory: true)
         let home = root.appendingPathComponent("home", isDirectory: true)
         let stateDir = root.appendingPathComponent("state", isDirectory: true)
         defer { try? FileManager().removeItem(at: root) }
 
-        try await TestIsolation.withEnvValues([
+        try await self.withLockedEnv([
             "OPENCLAW_HOME": home.path,
             "OPENCLAW_STATE_DIR": stateDir.path,
         ]) {
@@ -147,4 +197,19 @@ struct ExecApprovalsStoreRefactorTests {
         }
         return identifier
     }
+
+    private static func seedCurrentApprovalsFile(in stateDir: URL) throws {
+        try FileManager().createDirectory(at: stateDir, withIntermediateDirectories: true)
+        let file = ExecApprovalsFile(
+            version: 1,
+            socket: ExecApprovalsSocketConfig(
+                path: stateDir.appendingPathComponent("exec-approvals.sock").path,
+                token: "test-token"),
+            defaults: nil,
+            agents: [:])
+        let encoder = JSONEncoder()
+        encoder.outputFormatting = [.prettyPrinted, .sortedKeys]
+        try encoder.encode(file)
+            .write(to: stateDir.appendingPathComponent("exec-approvals.json"))
+    }
 }

docs/.generated/config-baseline.sha256

@@ -1,4 +1,4 @@
-37b56008790612b8293930b6a29d74490e98daa90f954fca9d133fcc28645c4c  config-baseline.json
-75b64c2ea081369ba4306493313a8a4cd48b784145f92fed995e6b77a5df350d  config-baseline.core.json
-17d64c9799dfa239a49493413f1100bdd9237e9b67aaeae331a4604dbc227023  config-baseline.channel.json
-f9d1f50bfa8403891e76cd99dc1357cdece4a71e8ae18a39b190c2a14e6f97b0  config-baseline.plugin.json
+0485ba902d2afd89d2c41cde7180d0cec2900b2db6804b9f97d42b7d85cd3af5  config-baseline.json
+72bb80be618406f3337eaa2560d2559a35e49bd29576de8dd4a3aec1a6a94d92  config-baseline.core.json
+1218f5555541b61bd5ddcac6441f15061b44789e2471d4ffecbe3059777c55c1  config-baseline.channel.json
+a14ac4261e98403d1a7e047070e6f151938444e27382b860315bd0c74fda4861  config-baseline.plugin.json

docs/.generated/plugin-sdk-api-baseline.sha256

@@ -1,2 +1,2 @@
-2c783beea6b3cda3d79060739a923f9f39e7e8b5942123dd6b08a09143a587ca  plugin-sdk-api-baseline.json
-0b33af2cffb42abb46682fb71c8f214da220793f13d10a34d332e75ff99e8ce9  plugin-sdk-api-baseline.jsonl
+5500f1695f29e18f52f4c9c19fbf1093613a4e4596a3a32619fe7c9da68828fe  plugin-sdk-api-baseline.json
+0b58595cee12ac1330e0d5b07eb09954ad6959118b7d19d2abebbf5158b3f0a1  plugin-sdk-api-baseline.jsonl

docs/channels/broadcast-groups.md

@@ -161,17 +161,20 @@ Control how agents process messages:
   <Step title="Incoming message arrives">
     A WhatsApp group or DM message arrives.
   </Step>
-  <Step title="Broadcast check">
-    System checks if peer ID is in `broadcast`.
+  <Step title="Route and admission">
+    OpenClaw applies channel allowlists, group activation rules, and configured ACP binding ownership.
   </Step>
-  <Step title="If in broadcast list">
+  <Step title="Broadcast check">
+    If no configured ACP binding owns the route, OpenClaw checks whether the peer ID is in `broadcast`.
+  </Step>
+  <Step title="If broadcast applies">
     - All listed agents process the message.
     - Each agent has its own session key and isolated context.
     - Agents process in parallel (default) or sequentially.
 
   </Step>
-  <Step title="If not in broadcast list">
-    Normal routing applies (first matching binding).
+  <Step title="If broadcast does not apply">
+    OpenClaw dispatches the ordinary route or the configured ACP session route selected during routing.
   </Step>
 </Steps>
 
@@ -322,7 +325,7 @@ Broadcast groups work alongside existing routing:
 - `GROUP_B`: agent1 AND agent2 respond (broadcast).
 
 <Note>
-**Precedence:** `broadcast` takes priority over `bindings`.
+**Precedence:** `broadcast` takes priority over ordinary route bindings. Configured ACP bindings (`bindings[].type="acp"`) are exclusive: when one matches, OpenClaw dispatches to the configured ACP session instead of fan-out broadcast.
 </Note>
 
 ## Troubleshooting
@@ -343,9 +346,9 @@ Broadcast groups work alongside existing routing:
 
   </Accordion>
   <Accordion title="Only one agent responding">
-    **Cause:** Peer ID might be in `bindings` but not `broadcast`.
+    **Cause:** Peer ID might be in ordinary route bindings but not `broadcast`, or it might match an exclusive configured ACP binding.
 
-    **Fix:** Add to broadcast config or remove from bindings.
+    **Fix:** Add ordinary route-bound peers to broadcast config, or remove/change the configured ACP binding if fan-out broadcast is desired.
 
   </Accordion>
   <Accordion title="Performance issues">

docs/channels/feishu.md

@@ -416,7 +416,9 @@ Enable `dynamicAgentCreation` to automatically create **isolated agent instances
 This is essential for public bots where you want each user to have their own private AI assistant experience.
 
 <Note>
-**Account limitation**: `dynamicAgentCreation` currently works with the **default Feishu account only**. Named/multi-account setups are not yet fully supported — dynamic bindings are created without `accountId`, so messages to named accounts may still route to `agent:main`. Track progress in [Issue #42837](https://github.com/openclaw/openclaw/issues/42837).
+Dynamic bindings include the normalized Feishu `accountId`, so default and named accounts route each sender to the correct dynamic agent.
+
+If a named account created an unscoped dynamic agent on an older release, that legacy agent still counts toward `maxAgents`. Confirm that it is not used by the default account before removing it, or temporarily increase `maxAgents`; OpenClaw cannot safely infer which account owns ambiguous legacy state.
 </Note>
 
 ### Quick setup
@@ -447,7 +449,7 @@ This is essential for public bots where you want each user to have their own pri
 
 When a new user sends their first DM:
 
-1. The channel generates a unique `agentId` = `feishu-{user_open_id}`
+1. The channel generates a unique `agentId`: `feishu-{user_open_id}` for the default account, or a bounded account-prefixed identity digest for a named account
 2. Creates a new workspace at `workspaceTemplate` path
 3. Registers the agent and creates a binding for this user
 4. The workspace helper ensures bootstrap files (`AGENTS.md`, `SOUL.md`, `USER.md`, etc.) on first access
@@ -464,22 +466,23 @@ When a new user sends their first DM:
 
 Template variables:
 
-- `{agentId}` - the generated agent ID (e.g., `feishu-ou_xxxxxx`)
+- `{agentId}` - the generated agent ID (e.g., `feishu-ou_xxxxxx` or `feishu-support-<identity_digest>`)
 - `{userId}` - the sender's Feishu open_id (e.g., `ou_xxxxxx`)
 
 ### Session scope
 
 `session.dmScope` controls how direct messages are mapped to agent sessions. This is a **global setting** that affects all channels.
 
-| Value                | Behavior                                                  | Best for                                                           |
-| -------------------- | --------------------------------------------------------- | ------------------------------------------------------------------ |
-| `"main"`             | Each user's DM maps to their agent's main session         | Single-user bots where you want `USER.md` / `SOUL.md` to auto-load |
-| `"per-channel-peer"` | Each (channel + user) combination gets a separate session | Public multi-user bots needing stronger isolation                  |
+| Value                        | Behavior                                                            | Best for                                                           |
+| ---------------------------- | ------------------------------------------------------------------- | ------------------------------------------------------------------ |
+| `"main"`                     | Each user's DM maps to their agent's main session                   | Single-user bots where you want `USER.md` / `SOUL.md` to auto-load |
+| `"per-channel-peer"`         | Each (channel + user) combination gets a separate session           | Public multi-user bots needing stronger isolation                  |
+| `"per-account-channel-peer"` | Each (account + channel + user) combination gets a separate session | Multi-account bots needing account-level session isolation         |
 
 **Tradeoff**: Using `"main"` enables automatic bootstrap file loading (`USER.md`, `SOUL.md`, `MEMORY.md`), but means all DMs across all channels share the same session key pattern. For public multi-user bots where isolation matters more than bootstrap auto-loading, consider `"per-channel-peer"` and manage bootstrap files manually.
 
 <Note>
-`"per-account-channel-peer"` is not recommended with `dynamicAgentCreation` because dynamic bindings are created without `accountId`. Use it only with manual bindings.
+Use `"per-account-channel-peer"` when named Feishu accounts should keep separate sessions for the same sender. Dynamic bindings preserve the account scope.
 </Note>
 
 ```json5

docs/channels/groups.md

@@ -586,7 +586,7 @@ Group inbound payloads set:
 - `WasMentioned` (mention gating result)
 - Telegram forum topics also include `MessageThreadId` and `IsForum`.
 
-The agent system prompt includes a group intro on the first turn of a new group session. It reminds the model to respond like a human, avoid Markdown tables, minimize empty lines and follow normal chat spacing, and avoid typing literal `\n` sequences. Channel-sourced group names and participant labels are rendered as fenced untrusted metadata, not inline system instructions.
+The agent system prompt includes a group intro on the first turn of a new group session. It reminds the model to respond like a human, minimize empty lines and follow normal chat spacing, and avoid typing literal `\n` sequences. Non-Telegram groups also discourage Markdown tables; Telegram rich-text guidance comes from the Telegram channel prompt. Channel-sourced group names and participant labels are rendered as fenced untrusted metadata, not inline system instructions.
 
 ## iMessage specifics
 

docs/channels/telegram.md

@@ -311,7 +311,6 @@ curl "https://api.telegram.org/bot<bot_token>/getUpdates"
 
     - direct chats: preview message + `editMessageText`
     - groups/topics: preview message + `editMessageText`
-    - direct-chat tool progress: optional native `sendMessageDraft` status preview when enabled and supported
 
     Requirement:
 
@@ -320,29 +319,10 @@ curl "https://api.telegram.org/bot<bot_token>/getUpdates"
     - `streaming.preview.toolProgress` controls whether tool/progress updates reuse the same edited preview message (default: `true` when preview streaming is active)
     - `streaming.preview.commandText` controls command/exec detail inside those tool-progress lines: `raw` (default, preserves released behavior) or `status` (tool label only)
     - `streaming.progress.commentary` (default: `false`) opts into assistant commentary/preamble text in the temporary progress draft
-    - legacy `channels.telegram.streamMode` and boolean `streaming` values are detected; run `openclaw doctor --fix` to migrate them to `channels.telegram.streaming.mode`
+    - legacy `channels.telegram.streamMode`, boolean `streaming` values, and retired native draft preview keys are detected; run `openclaw doctor --fix` to migrate them to current streaming config
 
     Tool-progress preview updates are the short status lines shown while tools run, for example command execution, file reads, planning updates, patch summaries, or Codex preamble/commentary text in Codex app-server mode. Telegram keeps these enabled by default to match released OpenClaw behavior from `v2026.4.22` and later.
 
-    Direct chats can use native Telegram drafts for these tool-progress lines without persisting tool chatter into chat history. Native drafts stop before answer text starts; final answers stay on the normal persistent delivery path. This lane is off by default and should be gated to trusted DM IDs first:
-
-    ```json
-    {
-      "channels": {
-        "telegram": {
-          "streaming": {
-            "mode": "partial",
-            "preview": {
-              "toolProgress": true,
-              "nativeToolProgress": true,
-              "nativeToolProgressAllowFrom": ["123456789"]
-            }
-          }
-        }
-      }
-    }
-    ```
-
     To keep the edited preview for answer text but hide tool-progress lines, set:
 
     ```json
@@ -420,14 +400,16 @@ curl "https://api.telegram.org/bot<bot_token>/getUpdates"
 
   </Accordion>
 
-  <Accordion title="Formatting and HTML fallback">
-    Outbound text uses Telegram `parse_mode: "HTML"`.
+  <Accordion title="Rich message formatting">
+    Outbound text uses Telegram rich messages.
 
-    - Markdown-ish text is rendered to Telegram-safe HTML.
-    - Supported Telegram HTML tags are preserved; unsupported HTML is escaped.
-    - If Telegram rejects parsed HTML, OpenClaw retries as plain text.
+    - Markdown text is sent as rich Markdown without converting it to HTML.
+    - Explicit HTML payloads are sent as rich HTML.
+    - Media captions still use Telegram HTML captions because rich messages do not replace captions.
 
-    Link previews are enabled by default and can be disabled with `channels.telegram.linkPreview: false`.
+    Long rich text is split automatically across Telegram's rich text and rich block limits. Tables over Telegram's column limit are sent as code blocks.
+
+    Link previews are enabled by default. `channels.telegram.linkPreview: false` skips automatic entity detection for rich text.
 
   </Accordion>
 

docs/channels/whatsapp.md

@@ -319,6 +319,40 @@ content and identifiers.
   </Tab>
 </Tabs>
 
+## Configured ACP bindings
+
+WhatsApp supports persistent ACP bindings with top-level `bindings[]` entries:
+
+```json5
+{
+  bindings: [
+    {
+      type: "acp",
+      agentId: "codex",
+      match: {
+        channel: "whatsapp",
+        accountId: "work",
+        peer: { kind: "direct", id: "+15555550123" },
+      },
+    },
+    {
+      type: "acp",
+      agentId: "codex",
+      match: {
+        channel: "whatsapp",
+        accountId: "work",
+        peer: { kind: "group", id: "120363424282127706@g.us" },
+      },
+    },
+  ],
+}
+```
+
+- Direct chats match E.164 numbers such as `+15555550123`.
+- Groups match WhatsApp group JIDs such as `120363424282127706@g.us`.
+- Group allowlists, sender policy, and mention or activation gating run before OpenClaw ensures the configured ACP session exists.
+- A matched configured ACP binding owns the route. WhatsApp broadcast groups do not fan out that turn to ordinary WhatsApp sessions.
+
 ## Personal-number and self-chat behavior
 
 When the linked self number is also present in `allowFrom`, WhatsApp self-chat safeguards activate:

docs/ci.md

@@ -200,13 +200,19 @@ from `release/YYYY.M.PATCH` or `main` after the release tag exists and after the
 OpenClaw npm preflight has succeeded. It verifies `pnpm plugins:sync:check`,
 dispatches `Plugin NPM Release` for all publishable plugin packages, dispatches
 `Plugin ClawHub Release` for the same release SHA, and only then dispatches
-`OpenClaw NPM Release` with the saved `preflight_run_id`.
+`OpenClaw NPM Release` with the saved `preflight_run_id`. Stable publish also
+requires an exact `windows_node_tag`; the workflow verifies the Windows source
+release and compares its x64/ARM64 installers with the candidate-approved
+`windows_node_installer_digests` input before any publish child, then promotes
+and verifies those same pinned installer digests plus the exact companion asset
+and checksum contract before publishing the GitHub release draft.
 
 ```bash
 gh workflow run openclaw-release-publish.yml \
   --ref release/YYYY.M.PATCH \
   -f tag=vYYYY.M.PATCH-beta.N \
   -f preflight_run_id=<successful-openclaw-npm-preflight-run-id> \
+  -f full_release_validation_run_id=<successful-full-release-validation-run-id> \
   -f npm_dist_tag=beta
 ```
 
@@ -452,7 +458,7 @@ For normal PRs, follow scoped CI/check evidence instead of treating parity as a
 
 The `CodeQL` workflow is intentionally a narrow first-pass security scanner, not the full repository sweep. Daily, manual, and non-draft pull request guard runs scan Actions workflow code plus the highest-risk JavaScript/TypeScript surfaces with high-confidence security queries filtered to high/critical `security-severity`.
 
-The pull request guard stays light: it only starts for changes under `.github/actions`, `.github/codeql`, `.github/workflows`, `packages`, `scripts`, `src`, or process-owning bundled plugin runtime paths, and it runs the same high-confidence security matrix as the scheduled workflow. Android and macOS CodeQL stay out of PR defaults.
+The pull request guard stays light: it only starts for changes under `.github/actions`, `.github/codeql`, `.github/workflows`, `packages`, or `src`, and it runs the same high-confidence security matrix as the scheduled workflow. Android and macOS CodeQL stay out of PR defaults.
 
 ### Security categories
 
@@ -462,7 +468,6 @@ The pull request guard stays light: it only starts for changes under `.github/ac
 | `/codeql-security-high/channel-runtime-boundary`  | Core channel implementation contracts plus the channel plugin runtime, gateway, Plugin SDK, secrets, audit touchpoints              |
 | `/codeql-security-high/network-ssrf-boundary`     | Core SSRF, IP parsing, network guard, web-fetch, and Plugin SDK SSRF policy surfaces                                                |
 | `/codeql-security-high/mcp-process-tool-boundary` | MCP servers, process execution helpers, outbound delivery, and agent tool-execution gates                                           |
-| `/codeql-security-high/process-exec-boundary`     | Local shell, process spawn helpers, subprocess-owning bundled plugin runtimes, and workflow script glue                             |
 | `/codeql-security-high/plugin-trust-boundary`     | Plugin install, loader, manifest, registry, package-manager install, source-loading, and Plugin SDK package contract trust surfaces |
 
 ### Platform-specific security shards

docs/cli/browser.md

@@ -174,7 +174,22 @@ Notes:
   or `--element`.
 - `existing-session` / `user` profiles support page screenshots and `--ref`
   screenshots from snapshot output, but not CSS `--element` screenshots.
-- `--labels` overlays current snapshot refs on the screenshot.
+- `--labels` overlays current snapshot refs on the screenshot. On
+  Playwright-backed profiles, it works with `--full-page` (full-page label
+  overlay), `--ref` (element-clip label overlay by ARIA ref), and `--element`
+  (element-clip label overlay by CSS selector); in element-clip modes, labels
+  are projected relative to the element. The response also includes an
+  `annotations` array with each ref's bounding box. Each item has `ref`,
+  `number`, `role`, optional `name`, and `box: {x, y, width, height}`;
+  coordinates are in the captured image's space (viewport / fullpage /
+  element-relative). The field is omitted when empty.
+  `existing-session` profiles render a chrome-mcp overlay on page screenshots
+  but do not use the Playwright projection helper and do not include
+  `annotations`; CSS `--element` screenshots are unsupported there. Without
+  Playwright or chrome-mcp, labeled screenshots are not available. Prior
+  releases ignored `--full-page`, `--ref`, and `--element` on labeled
+  Playwright screenshots and always returned a viewport capture; labeled
+  screenshots now honor those scopes.
 - `snapshot --urls` appends discovered link destinations to AI snapshots so
   agents can choose direct navigation targets instead of guessing from link
   text alone.

docs/cli/onboard.md

@@ -182,7 +182,10 @@ Interactive onboarding behavior with reference mode:
 ### Non-interactive Z.AI endpoint choices
 
 <Note>
-`--auth-choice zai-api-key` auto-detects the best Z.AI endpoint for your key (prefers the general API with `zai/glm-5.1`). If you specifically want the GLM Coding Plan endpoints, pick `zai-coding-global` or `zai-coding-cn`.
+`--auth-choice zai-api-key` auto-detects the best Z.AI endpoint and model for
+your key. Coding Plan endpoints prefer `zai/glm-5.2`; general API endpoints use
+`zai/glm-5.1`. To force a Coding Plan endpoint, pick `zai-coding-global` or
+`zai-coding-cn`.
 </Note>
 
 ```bash

docs/cli/plugins.md

@@ -159,7 +159,7 @@ is available, then fall back to `latest`.
   <Accordion title="--dangerously-force-unsafe-install">
     `--dangerously-force-unsafe-install` is deprecated and is now a no-op. OpenClaw no longer runs built-in install-time dangerous-code blocking for plugin installs.
 
-    Use the shared operator-owned `security.installPolicy` surface when host-specific install policy is required. Plugin `before_install` hooks and `security.installPolicy` can still block installs.
+    Use the shared operator-owned `security.installPolicy` surface when host-specific install policy is required. Plugin `before_install` hooks are plugin-runtime lifecycle hooks and are not the primary policy boundary for CLI installs.
 
     If a plugin you published on ClawHub is hidden or blocked by a registry scan, use the publisher steps in [ClawHub publishing](/clawhub/publishing). `--dangerously-force-unsafe-install` does not ask ClawHub to rescan the plugin or make a blocked release public.
 
@@ -405,7 +405,7 @@ Updates apply to tracked plugin installs in the managed plugin index and tracked
 
   </Accordion>
   <Accordion title="--dangerously-force-unsafe-install on update">
-    `--dangerously-force-unsafe-install` is also accepted on `plugins update` for compatibility, but it is deprecated and no longer changes plugin update behavior. Operator `security.installPolicy` and plugin `before_install` hooks can still block updates.
+    `--dangerously-force-unsafe-install` is also accepted on `plugins update` for compatibility, but it is deprecated and no longer changes plugin update behavior. Operator `security.installPolicy` can still block updates; plugin `before_install` hooks only apply in processes where plugin hooks are loaded.
   </Accordion>
 </AccordionGroup>
 

docs/concepts/active-memory.md

@@ -479,6 +479,9 @@ names that plugin registers. Active Memory lists those tools in the recall
 prompt and passes the same list to the embedded sub-agent. If none of the
 configured tools are available, or the memory sub-agent fails, Active Memory
 skips recall for that turn and the main reply continues without memory context.
+For custom recall tools, non-empty model-visible tool output counts as recall
+evidence unless structured result fields explicitly report an empty result or
+failure.
 `toolsAllow` only accepts concrete memory tool names. Wildcards, `group:*`
 entries, and core agent tools such as `read`, `exec`, `message`, and
 `web_search` are ignored before the hidden memory sub-agent starts.
@@ -743,7 +746,11 @@ Before v2026.5.2 the plugin silently extended your configured `timeoutMs` by an
 extra 30000 ms during cold-start so model warm-up, embedding-index load, and
 the first recall could share one larger budget. v2026.5.2 moved that grace
 behind an explicit `setupGraceTimeoutMs` config — your configured `timeoutMs`
-is now the budget by default, unless you opt in.
+is now the recall-work budget by default, unless you opt in. The blocking hook
+uses two bounded phases around that budget: up to 1500 ms for session/config
+preflight before recall starts, then a separate fixed 1500 ms for abort
+settlement and transcript recovery after recall work stops. Neither allowance
+extends model or tool execution.
 
 If you upgraded from v2026.4.x and you set `timeoutMs` to a value tuned for the
 old implicit-grace world (the recommended starter `timeoutMs: 15000` is one
@@ -765,14 +772,16 @@ outer watchdog budgets back to the pre-v5.2 effective values:
 }
 ```
 
-Per the v2026.5.2 changelog: _"use the configured recall timeout as the
-blocking prompt-build hook budget by default and move cold-start setup grace
-behind explicit `setupGraceTimeoutMs` config, so the plugin no longer silently
-extends 15000 ms configs to 45000 ms on the main lane."_
+The v2026.5.2 change removed the old implicit 30000 ms cold-start extension.
+Beyond the configured recall-work budget, the hook can use up to 1500 ms for
+preflight and another 1500 ms for post-recall completion. Its worst-case
+blocking time is therefore `timeoutMs + setupGraceTimeoutMs + 3000` ms.
 
 The embedded recall runner uses the same effective timeout budget, so
 `setupGraceTimeoutMs` covers both the outer prompt-build watchdog and the inner
-blocking recall run.
+blocking recall run. The preflight cap covers session/config checks before that
+budget begins. The post-recall allowance lets the outer hook settle abort
+cleanup and read any final transcript state.
 
 For resource-tight gateways where cold-start latency is a known trade-off,
 lower values (5000–15000 ms) work too — the trade-off is a higher chance of

docs/concepts/agent-loop.md

@@ -97,7 +97,7 @@ These run inside the agent loop or gateway pipeline:
 - **`agent_end`**: inspect the final message list and run metadata after completion.
 - **`before_compaction` / `after_compaction`**: observe or annotate compaction cycles.
 - **`before_tool_call` / `after_tool_call`**: intercept tool params/results.
-- **`before_install`**: inspect install context and optionally block skill or plugin installs after operator install policy runs.
+- **`before_install`**: inspect staged skill or plugin install material after operator install policy runs, when plugin hooks are loaded in the current OpenClaw process.
 - **`tool_result_persist`**: synchronously transform tool results before they are written to an OpenClaw-owned session transcript.
 - **`message_received` / `message_sending` / `message_sent`**: inbound + outbound message hooks.
 - **`session_start` / `session_end`**: session lifecycle boundaries.
@@ -109,6 +109,7 @@ Hook decision rules for outbound/tool guards:
 - `before_tool_call`: `{ block: false }` is a no-op and does not clear a prior block.
 - `before_install`: `{ block: true }` is terminal and stops lower-priority handlers.
 - `before_install`: `{ block: false }` is a no-op and does not clear a prior block.
+- Use `security.installPolicy`, not `before_install`, for operator-owned install allow/block decisions that must cover CLI install and update paths.
 - `message_sending`: `{ cancel: true }` is terminal and stops lower-priority handlers.
 - `message_sending`: `{ cancel: false }` is a no-op and does not clear a prior cancel.
 

docs/concepts/mantis.md

@@ -247,12 +247,13 @@ of only a bot-to-bot Slack transcript.
 evidence pipeline. It checks out the trusted candidate ref in a separate
 worktree, runs `pnpm openclaw qa telegram --credential-source convex
 --credential-role ci`, writes a `mantis-evidence.json` manifest from the
-Telegram QA summary and observed-message artifact, renders the redacted
-transcript HTML through a Crabbox desktop browser, generates a motion-trimmed GIF
-with `crabbox media preview`, and posts the inline PR evidence comment when a PR
-number is available. This lane is transcript-visual rather than logged-in
-Telegram Web proof: the Telegram Bot API gives stable live message evidence, but
-Telegram Web login state is not required for normal Mantis automation.
+Telegram QA summary, `qa-evidence.json`, and report artifacts, renders the
+redacted evidence HTML through a Crabbox desktop browser, generates a
+motion-trimmed GIF with `crabbox media preview`, and posts the inline PR
+evidence comment when a PR number is available. This lane is QA-evidence visual
+rather than logged-in Telegram Web proof: the Telegram Bot API gives stable live
+message evidence, but Telegram Web login state is not required for normal Mantis
+automation.
 
 `Mantis Telegram Desktop Proof` is the agentic native Telegram Desktop
 before/after wrapper. A maintainer can trigger it from a PR comment with
@@ -494,8 +495,8 @@ zero:
 
 - `pnpm openclaw qa discord` already runs a live Discord lane with driver and
   SUT bots.
-- The live transport runner already writes reports and observed-message
-  artifacts under `.artifacts/qa-e2e/`.
+- The live transport runner already writes reports, QA evidence, and
+  transport-specific artifacts under `.artifacts/qa-e2e/`.
 - Convex credential leases already provide exclusive access to shared live
   transport credentials.
 - The browser control service already supports screenshots, snapshots,

docs/concepts/model-providers.md

@@ -264,7 +264,7 @@ Gemini CLI JSON replies are parsed from `response`; usage falls back to `stats`,
 
 - Provider: `zai`
 - Auth: `ZAI_API_KEY`
-- Example model: `zai/glm-5.1`
+- Example model: `zai/glm-5.2`
 - CLI: `openclaw onboard --auth-choice zai-api-key`
   - Model refs use the canonical `zai/*` provider ID.
   - `zai-api-key` auto-detects the matching Z.AI endpoint; `zai-coding-global`, `zai-coding-cn`, `zai-global`, and `zai-cn` force a specific surface

docs/concepts/qa-e2e-automation.md

@@ -318,17 +318,17 @@ Matrix has a [dedicated page](/concepts/qa-matrix) because of its scenario count
 
 These lanes register through `extensions/qa-lab/src/live-transports/shared/live-transport-cli.ts` and accept the same flags:
 
-| Flag                                  | Default                                            | Description                                                                                                           |
-| ------------------------------------- | -------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------- |
-| `--scenario <id>`                     | -                                                  | Run only this scenario. Repeatable.                                                                                   |
-| `--output-dir <path>`                 | `<repo>/.artifacts/qa-e2e/<transport>-<timestamp>` | Where reports/summary/observed messages and the output log are written. Relative paths resolve against `--repo-root`. |
-| `--repo-root <path>`                  | `process.cwd()`                                    | Repository root when invoking from a neutral cwd.                                                                     |
-| `--sut-account <id>`                  | `sut`                                              | Temporary account id inside the QA gateway config.                                                                    |
-| `--provider-mode <mode>`              | `live-frontier`                                    | `mock-openai` or `live-frontier` (legacy `live-openai` still works).                                                  |
-| `--model <ref>` / `--alt-model <ref>` | provider default                                   | Primary/alternate model refs.                                                                                         |
-| `--fast`                              | off                                                | Provider fast mode where supported.                                                                                   |
-| `--credential-source <env\|convex>`   | `env`                                              | See [Convex credential pool](#convex-credential-pool).                                                                |
-| `--credential-role <maintainer\|ci>`  | `ci` in CI, `maintainer` otherwise                 | Role used when `--credential-source convex`.                                                                          |
+| Flag                                  | Default                                            | Description                                                                                                                                     |
+| ------------------------------------- | -------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
+| `--scenario <id>`                     | -                                                  | Run only this scenario. Repeatable.                                                                                                             |
+| `--output-dir <path>`                 | `<repo>/.artifacts/qa-e2e/<transport>-<timestamp>` | Where reports, summaries, evidence, transport-specific artifacts, and the output log are written. Relative paths resolve against `--repo-root`. |
+| `--repo-root <path>`                  | `process.cwd()`                                    | Repository root when invoking from a neutral cwd.                                                                                               |
+| `--sut-account <id>`                  | `sut`                                              | Temporary account id inside the QA gateway config.                                                                                              |
+| `--provider-mode <mode>`              | `live-frontier`                                    | `mock-openai` or `live-frontier` (legacy `live-openai` still works).                                                                            |
+| `--model <ref>` / `--alt-model <ref>` | provider default                                   | Primary/alternate model refs.                                                                                                                   |
+| `--fast`                              | off                                                | Provider fast mode where supported.                                                                                                             |
+| `--credential-source <env\|convex>`   | `env`                                              | See [Convex credential pool](#convex-credential-pool).                                                                                          |
+| `--credential-role <maintainer\|ci>`  | `ci` in CI, `maintainer` otherwise                 | Role used when `--credential-source convex`.                                                                                                    |
 
 Each lane exits non-zero on any failed scenario. `--allow-failures` writes artifacts without setting a failing exit code.
 
@@ -346,10 +346,6 @@ Required env when `--credential-source env`:
 - `OPENCLAW_QA_TELEGRAM_DRIVER_BOT_TOKEN`
 - `OPENCLAW_QA_TELEGRAM_SUT_BOT_TOKEN`
 
-Optional:
-
-- `OPENCLAW_QA_TELEGRAM_CAPTURE_CONTENT=1` keeps message bodies in observed-message artifacts (default redacts).
-
 Scenarios (`extensions/qa-lab/src/live-transports/telegram/telegram-live.runtime.ts`):
 
 - `telegram-canary`
@@ -375,26 +371,26 @@ Output artifacts:
 
 - `telegram-qa-report.md`
 - `qa-evidence.json` - evidence entries for the live transport checks, including profile, coverage, provider, channel, artifacts, result, and RTT fields.
-- `telegram-qa-observed-messages.json` - bodies redacted unless `OPENCLAW_QA_TELEGRAM_CAPTURE_CONTENT=1`.
 
-Package RTT comparison uses the same Telegram credential contract while keeping
-its RTT sample controls on the RTT harness path:
+Package Telegram runs use the same Telegram credential contract. Repeated RTT
+measurement is part of the normal package Telegram live lane; the RTT
+distribution is folded into `qa-evidence.json` under `result.timing` for the
+selected RTT check.
 
 ```bash
-pnpm rtt openclaw@beta \
-  --credential-source convex \
-  --credential-role maintainer \
-  --samples 20 \
-  --sample-timeout-ms 30000
+OPENCLAW_QA_CREDENTIAL_SOURCE=convex \
+pnpm test:docker:npm-telegram-live
 ```
 
-When `--credential-source convex` is set, the RTT Docker wrapper leases a
-`kind: "telegram"` credential, exports the leased group/driver/SUT bot env into
-the installed-package run, heartbeats the lease, and releases it on shutdown.
-`--samples` and `--sample-timeout-ms` still feed
-`OPENCLAW_NPM_TELEGRAM_WARM_SAMPLES` and
-`OPENCLAW_NPM_TELEGRAM_SAMPLE_TIMEOUT_MS`, so `result.json` remains comparable
-across env-backed and Convex-backed RTT runs.
+When `OPENCLAW_QA_CREDENTIAL_SOURCE=convex` is set, the package live wrapper
+leases a `kind: "telegram"` credential, exports the leased group/driver/SUT bot
+env into the installed-package run, heartbeats the lease, and releases it on
+shutdown. The package wrapper defaults to 20 RTT checks of
+`telegram-mentioned-message-reply`, a 30s RTT timeout, and Convex role
+`maintainer` outside CI when Convex is selected. Override
+`OPENCLAW_NPM_TELEGRAM_RTT_SAMPLES`, `OPENCLAW_NPM_TELEGRAM_RTT_TIMEOUT_MS`,
+or `OPENCLAW_NPM_TELEGRAM_RTT_MAX_FAILURES` to tune RTT measurement without
+creating a separate RTT command or Telegram-specific summary format.
 
 ### Discord QA
 

docs/concepts/usage-tracking.md

@@ -32,8 +32,13 @@ title: "Usage tracking"
 
 ## Custom `/usage full` footer
 
-Set `messages.usageTemplate` to customize the per-response `/usage full`
-footer. The value can be an inline template object or a JSON file path:
+`/usage full` shows a built-in compact footer with model, reasoning, fast/slow,
+context window, turn tokens, cache, and cost when those fields are available. No
+template file is required.
+
+`messages.usageTemplate` is only for advanced custom layouts. The value is a
+JSON file path (supports `~`) or an inline object, and it replaces the built-in
+footer when valid:
 
 ```json
 {
@@ -43,9 +48,182 @@ footer. The value can be an inline template object or a JSON file path:
 }
 ```
 
-Templates read the `openclaw.usageLine.v1` contract and can use `scales`,
-`aliases`, and `output.surfaces` to render channel-specific footers. Missing,
-unreadable, invalid, or empty templates fall back to the built-in usage line.
+Missing or empty templates fall back to the built-in footer quietly. Unreadable
+or invalid configured templates also fall back to the built-in footer and emit an
+operator warning.
+
+Start custom templates from the built-in shape, then edit the parts you want to
+change:
+
+```jsonc
+{
+  "schema": "openclaw.usageBar.v1",
+  "scales": {
+    "braille": "⠐⡀⡄⡆⡇⣇⣧⣷⣿",
+    "block": "░▏▎▍▌▋▊▉█",
+    "shade": "░▒▓█",
+    "moon": "🌑🌘🌗🌖🌕",
+    "level": "▁▂▃▄▅▆▇█",
+    "weather": ["🥶", "☁️", "🌥", "⛅️", "🌤", "☀️"],
+    "plants": ["🪾", "🍂", "🌱", "☘️", "🍀", "🌿"],
+    "moons6": ["🌑", "🌚", "🌘", "🌗", "🌖", "🌝"],
+  },
+  "aliases": {
+    "models": {
+      "claude-opus-4-6": "opus46",
+      "claude-opus-4-8": "opus48",
+      "claude-sonnet-4-6": "sonnet46",
+      "claude-haiku-4-5": "haiku45",
+      "gpt-5.5": "gpt5.5",
+    },
+    "reasoning": {
+      "off": "🌑",
+      "minimal": "🌚",
+      "low": "🌘",
+      "medium": "🌗",
+      "high": "🌕",
+      "xhigh": "🌝",
+    },
+  },
+  "output": {
+    "sep": "",
+    "default": [
+      { "text": "{model.provider}{identity.emoji|🤖} {model.display_name|alias:models}" },
+      { "map": "model.is_fallback", "cases": { "true": " 🔄" } },
+      { "map": "model.is_override", "cases": { "true": " 📌" } },
+      { "when": "model.reasoning", "text": " {model.reasoning|alias:reasoning}" },
+      { "map": "state.fast_mode", "cases": { "true": " ⚡", "false": " 🐌" } },
+      {
+        "when": "context.max_tokens",
+        "text": " | 📚 [{context.pct_used|meter:5:braille}]{context.max_tokens|num}",
+      },
+      {
+        "when": "usage.has_split_tokens",
+        "text": " ↕️ {usage.input_tokens|num|?}/{usage.output_tokens|num|?}",
+      },
+      { "when": "usage.has_total_only_tokens", "text": " ↕️ {usage.total_tokens|num}" },
+      { "when": "usage.cache_hit_pct", "text": " 🗄 {usage.cache_hit_pct|pct}" },
+      { "when": "cost.turn_usd", "text": " 💰{cost.turn_usd|fixed:4}" },
+    ],
+    "surfaces": {
+      "discord": [
+        { "text": "-# -\n" },
+        { "text": "-# {model.provider}{identity.emoji|🤖} {model.display_name|alias:models}" },
+        { "map": "model.is_fallback", "cases": { "true": "🔄" } },
+        { "map": "model.is_override", "cases": { "true": "📌" } },
+        { "when": "model.reasoning", "text": " {model.reasoning|alias:reasoning}" },
+        { "map": "state.fast_mode", "cases": { "true": " ⚡️", "false": " 🐌" } },
+        {
+          "when": "context.max_tokens",
+          "text": " | 📚 [{context.pct_used|meter:5:braille}]{context.max_tokens|num}",
+        },
+        {
+          "when": "usage.has_split_tokens",
+          "text": " ↕️ {usage.input_tokens|num|?}/{usage.output_tokens|num|?}",
+        },
+        { "when": "usage.has_total_only_tokens", "text": " ↕️ {usage.total_tokens|num}" },
+        { "when": "usage.cache_hit_pct", "text": " 🗄 {usage.cache_hit_pct|pct}" },
+        { "when": "cost.turn_usd", "text": " 💰{cost.turn_usd|fixed:4}" },
+      ],
+    },
+  },
+}
+```
+
+### Shape
+
+```jsonc
+{
+  "schema": "openclaw.usageBar.v1",
+  "scales": { "<name>": "low-to-high glyphs" }, // string (1 glyph/char) or array
+  "aliases": { "<table>": { "<value>": "<label>" } },
+  "output": {
+    "sep": "", // joins surviving pieces
+    "default": [
+      /* pieces */
+    ], // fallback for any surface
+    "surfaces": {
+      "discord": [
+        /* pieces */
+      ],
+      "telegram": [
+        /* pieces */
+      ],
+    },
+  },
+}
+```
+
+Each surface is an ordered list of **pieces**; the engine renders each, drops
+empties, and joins survivors with `sep`. A surface with no entry uses
+`output.default`.
+
+### Contract Paths
+
+A piece reads values from the per-turn contract by dot-path. Absent values are
+empty (so a `when` guard or a `|fallback` keeps the piece clean).
+
+| Path                                                                                | Meaning                                |
+| ----------------------------------------------------------------------------------- | -------------------------------------- |
+| `surface`                                                                           | channel id (`discord`/`telegram`/etc.) |
+| `model.provider` / `model.display_name`                                             | provider id / model id                 |
+| `model.reasoning`                                                                   | effort (`off` through `xhigh`)         |
+| `model.is_fallback` / `model.is_override`                                           | bool: fallback used / model pinned     |
+| `state.fast_mode`                                                                   | bool: fast vs slow                     |
+| `context.max_tokens` / `context.pct_used`                                           | window budget / 0-100 used             |
+| `usage.input_tokens` / `usage.output_tokens` / `usage.total_tokens`                 | turn aggregate                         |
+| `usage.has_split_tokens` / `usage.has_total_only_tokens` / `usage.cache_hit_pct`    | token display guards and cache percent |
+| `usage.last.input_tokens` / `usage.last.output_tokens` / `usage.last.cache_hit_pct` | final model call only                  |
+| `cost.turn_usd`                                                                     | estimated turn cost                    |
+| `identity.name` / `identity.emoji`                                                  | agent name / chosen emoji              |
+
+(Provider rate-limit windows are **not** in this contract.)
+
+### Verbs
+
+Pipe a value through verbs left to right; a non-verb segment is the fallback.
+
+| Verb            | Effect                                | Example                           |
+| --------------- | ------------------------------------- | --------------------------------- |
+| `num`           | compact count                         | `272000 -> 272k`                  |
+| `fixed:N`       | N decimals (default 2)                | `0.0377`                          |
+| `dur`           | seconds to duration                   | `14820 -> 4h07m`                  |
+| `pct`           | append `%`                            | `96 -> 96%`                       |
+| `inv`           | `100 - x`                             | for used to remaining             |
+| `alias:TABLE`   | lookup in `aliases`, echo if unlisted | `medium -> 🌗`                    |
+| `meter:W:SCALE` | W-cell glyph bar over a 0-100 value   | `[⣿⣿⠐⠐⠐]` (`meter:1` = one glyph) |
+
+### Piece forms
+
+- `{ "text": "📚 {context.max_tokens|num}" }`: literal + interpolation.
+- `{ "when": "<path>", "text": "..." }`: render only if the path is truthy.
+- `{ "map": "<path>", "cases": { "true": "⚡", "false": "🐌" } }`: value to glyph.
+- `{ "each": "limits.windows", "item": "{label}" }`: iterate an array.
+
+### Example
+
+```jsonc
+{
+  "schema": "openclaw.usageBar.v1",
+  "scales": { "braille": "⠐⡀⡄⡆⡇⣇⣧⣷⣿" },
+  "aliases": { "reasoning": { "medium": "🌗", "high": "🌕" } },
+  "output": {
+    "surfaces": {
+      "discord": [
+        { "text": "{model.display_name}" },
+        { "when": "model.reasoning", "text": " {model.reasoning|alias:reasoning}" },
+        { "map": "state.fast_mode", "cases": { "true": " ⚡", "false": " 🐌" } },
+        {
+          "when": "context.max_tokens",
+          "text": " | 📚 [{context.pct_used|meter:5:braille}]{context.max_tokens|num}",
+        },
+      ],
+    },
+  },
+}
+```
+
+renders e.g. `claude-sonnet-4-6 🌗 🐌 | 📚 [⣿⣿⣿⣿⣧]272k`.
 
 ## Providers + credentials
 

docs/gateway/config-channels.md

@@ -130,6 +130,8 @@ WhatsApp runs through the gateway's web channel (Baileys Web). It starts automat
 }
 ```
 
+- Top-level `bindings[]` entries with `type: "acp"` configure persistent ACP bindings for WhatsApp DMs and groups. Use an E.164 direct number or WhatsApp group JID in `match.peer.id`. Field semantics are shared in [ACP Agents](/tools/acp-agents#persistent-channel-bindings).
+
 <Accordion title="Multi-account WhatsApp">
 
 ```json5

docs/gateway/config-tools.md

@@ -339,7 +339,7 @@ Configures inbound media understanding (image/audio/video):
 
     - `capabilities`: optional list (`image`, `audio`, `video`). Defaults: `openai`/`anthropic`/`minimax` → image, `google` → image+audio+video, `groq` → audio.
     - `prompt`, `maxChars`, `maxBytes`, `timeoutSeconds`, `language`: per-entry overrides.
-    - `tools.media.image.timeoutSeconds` and matching image model `timeoutSeconds` entries also apply when the agent calls the explicit `image` tool.
+    - `tools.media.image.timeoutSeconds` and matching image model `timeoutSeconds` entries also apply when the agent calls the explicit `image` tool. For image understanding, this timeout applies to the request itself and is not reduced by earlier preparation work.
     - Failures fall back to the next entry.
 
     Provider auth follows standard order: `auth-profiles.json` → env vars → `models.providers.*.apiKey`.

docs/help/testing-live.md

@@ -73,7 +73,7 @@ Live tests are split into two layers so we can isolate failures:
   - `pnpm test:live` (or `OPENCLAW_LIVE_TEST=1` if invoking Vitest directly)
 - Set `OPENCLAW_LIVE_MODELS=modern`, `small`, or `all` (alias for modern) to actually run this suite; otherwise it skips to keep `pnpm test:live` focused on gateway smoke
 - How to select models:
-  - `OPENCLAW_LIVE_MODELS=modern` to run the modern allowlist (Opus/Sonnet 4.6+, GPT-5.2 + Codex, Gemini 3, DeepSeek V4, GLM 4.7, MiniMax M3, Grok 4.3)
+  - `OPENCLAW_LIVE_MODELS=modern` to run the modern allowlist (Opus/Sonnet 4.6+, GPT-5.2 + Codex, Gemini 3, DeepSeek V4, GLM 5.1, MiniMax M3, Grok 4.3)
   - `OPENCLAW_LIVE_MODELS=small` to run the constrained small-model allowlist (Qwen 8B/9B local-compatible routes, Ollama Gemma, OpenRouter Qwen/GLM, and Z.AI GLM)
   - `OPENCLAW_LIVE_MODELS=all` is an alias for the modern allowlist
   - or `OPENCLAW_LIVE_MODELS="openai/gpt-5.5,anthropic/claude-opus-4-6,..."` (comma allowlist)
@@ -357,6 +357,9 @@ Narrow, explicit allowlists are fastest and least flaky:
 - Tool calling across several providers:
   - `OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.5,anthropic/claude-opus-4-6,google/gemini-3-flash-preview,deepseek/deepseek-v4-flash,zai/glm-5.1,minimax/MiniMax-M3" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
 
+- Z.AI Coding Plan GLM-5.2 direct smoke:
+  - `ZAI_CODING_LIVE_TEST=1 pnpm test:live src/agents/zai.live.test.ts`
+
 - Google focus (Gemini API key + Antigravity):
   - Gemini (API key): `OPENCLAW_LIVE_GATEWAY_MODELS="google/gemini-3-flash-preview" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
   - Antigravity (OAuth): `OPENCLAW_LIVE_GATEWAY_MODELS="google-antigravity/claude-opus-4-6-thinking,google-antigravity/gemini-3-pro-high" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
@@ -388,7 +391,7 @@ This is the "common models" run we expect to keep working:
 - Google (Gemini API): `google/gemini-3.1-pro-preview` and `google/gemini-3-flash-preview` (avoid older Gemini 2.x models)
 - Google (Antigravity): `google-antigravity/claude-opus-4-6-thinking` and `google-antigravity/gemini-3-flash`
 - DeepSeek: `deepseek/deepseek-v4-flash` and `deepseek/deepseek-v4-pro`
-- Z.AI (GLM): `zai/glm-5.1`
+- Z.AI (GLM): `zai/glm-5.1` (general API) or `zai/glm-5.2` (Coding Plan)
 - MiniMax: `minimax/MiniMax-M3`
 
 Run gateway smoke with tools + image:
@@ -402,7 +405,7 @@ Pick at least one per provider family:
 - Anthropic: `anthropic/claude-opus-4-6` (or `anthropic/claude-sonnet-4-6`)
 - Google: `google/gemini-3-flash-preview` (or `google/gemini-3.1-pro-preview`)
 - DeepSeek: `deepseek/deepseek-v4-flash`
-- Z.AI (GLM): `zai/glm-5.1`
+- Z.AI (GLM): `zai/glm-5.1` (general API) or `zai/glm-5.2` (Coding Plan)
 - MiniMax: `minimax/MiniMax-M3`
 
 Optional additional coverage (nice to have):

docs/help/testing.md

@@ -218,17 +218,27 @@ inside every shard.
     `OPENCLAW_NPM_TELEGRAM_PACKAGE_TGZ=/path/to/openclaw-current.tgz` or
     `OPENCLAW_CURRENT_PACKAGE_TGZ` to test a resolved local tarball instead of
     installing from the registry.
+  - Emits repeated RTT timing in `qa-evidence.json` by default with
+    `OPENCLAW_NPM_TELEGRAM_RTT_SAMPLES=20`. Override
+    `OPENCLAW_NPM_TELEGRAM_RTT_SAMPLES`,
+    `OPENCLAW_NPM_TELEGRAM_RTT_TIMEOUT_MS`, or
+    `OPENCLAW_NPM_TELEGRAM_RTT_MAX_FAILURES` to tune the RTT run.
+    `OPENCLAW_NPM_TELEGRAM_RTT_CHECKS` accepts a comma-separated list of
+    Telegram QA check IDs to sample; when unset, the default RTT-capable check
+    is `telegram-mentioned-message-reply`.
   - Uses the same Telegram env credentials or Convex credential source as
     `pnpm openclaw qa telegram`. For CI/release automation, set
     `OPENCLAW_NPM_TELEGRAM_CREDENTIAL_SOURCE=convex` plus
-    `OPENCLAW_QA_CONVEX_SITE_URL` and the role secret. If
+    `OPENCLAW_QA_CONVEX_SITE_URL` and a role secret. If
     `OPENCLAW_QA_CONVEX_SITE_URL` and a Convex role secret are present in CI,
     the Docker wrapper selects Convex automatically.
   - The wrapper validates Telegram or Convex credential env on the host before
     Docker build/install work. Set `OPENCLAW_NPM_TELEGRAM_SKIP_CREDENTIAL_PREFLIGHT=1`
     only when deliberately debugging pre-credential setup.
   - `OPENCLAW_NPM_TELEGRAM_CREDENTIAL_ROLE=ci|maintainer` overrides the shared
-    `OPENCLAW_QA_CREDENTIAL_ROLE` for this lane only.
+    `OPENCLAW_QA_CREDENTIAL_ROLE` for this lane only. When Convex credentials
+    are selected and no role is set, the wrapper uses `ci` in CI and
+    `maintainer` outside CI.
   - GitHub Actions exposes this lane as the manual maintainer workflow
     `NPM Telegram Beta E2E`. It does not run on merge. The workflow uses the
     `qa-live-shared` environment and Convex CI credential leases.
@@ -344,11 +354,11 @@ gh workflow run package-acceptance.yml --ref main \
     want artifacts without a failing exit code.
   - Requires two distinct bots in the same private group, with the SUT bot exposing a Telegram username.
   - For stable bot-to-bot observation, enable Bot-to-Bot Communication Mode in `@BotFather` for both bots and ensure the driver bot can observe group bot traffic.
-  - Writes a Telegram QA report, summary, and observed-messages artifact under `.artifacts/qa-e2e/...`. Replying scenarios include RTT from driver send request to observed SUT reply.
+  - Writes a Telegram QA report, summary, and `qa-evidence.json` under `.artifacts/qa-e2e/...`. Replying scenarios include RTT from driver send request to observed SUT reply.
 
 `Mantis Telegram Live` is the PR-evidence wrapper around this lane. It runs the
-candidate ref with Convex-leased Telegram credentials, renders the redacted
-observed-message transcript in a Crabbox desktop browser, records MP4 evidence,
+candidate ref with Convex-leased Telegram credentials, renders the redacted QA
+report/evidence bundle in a Crabbox desktop browser, records MP4 evidence,
 generates a motion-trimmed GIF, uploads the artifact bundle, and posts inline PR
 evidence through the Mantis GitHub App when `pr_number` is set. Maintainers can
 start it from the Actions UI through `Mantis Scenario` (`scenario_id:

docs/nodes/index.md

@@ -214,6 +214,59 @@ permission boundary. Dangerous plugin node commands still require explicit
 After a node changes its declared command list, reject the old device pairing
 and approve the new request so the gateway stores the updated command snapshot.
 
+## Config (`openclaw.json`)
+
+Node-related settings live under `gateway.nodes` and `tools.exec`:
+
+```json5
+{
+  gateway: {
+    nodes: {
+      // Auto-approve first-time node pairing from trusted networks (CIDR list).
+      // Disabled when unset. Only applies to first-time role:node requests
+      // with no requested scopes; does not auto-approve upgrades.
+      pairing: {
+        autoApproveCidrs: ["192.168.1.0/24"],
+      },
+      // Opt into dangerous/privacy-heavy node commands (camera.snap, etc.).
+      allowCommands: ["camera.snap", "screen.record"],
+      // Block exact command names even if defaults or allowCommands include them.
+      denyCommands: ["camera.clip"],
+    },
+  },
+  tools: {
+    exec: {
+      // Default exec host: "node" routes all exec calls to a paired node.
+      host: "node",
+      // Security mode for node exec: allow only approved/allowlisted commands.
+      security: "allowlist",
+      // Pin exec to a specific node (id or name). Omit to allow any node.
+      node: "build-node",
+    },
+  },
+}
+```
+
+Use exact node command names. `denyCommands` removes a command even when a
+platform default or `allowCommands` entry would otherwise allow it. See
+[Gateway configuration reference](/gateway/configuration-reference#gateway-field-details)
+for gateway node pairing and command-policy field details.
+
+Per-agent exec node override:
+
+```json5
+{
+  agents: {
+    list: [
+      {
+        id: "main",
+        tools: { exec: { node: "build-node" } },
+      },
+    ],
+  },
+}
+```
+
 ## Screenshots (canvas snapshots)
 
 If the node is showing the Canvas (WebView), `canvas.snapshot` returns `{ format, base64 }`.

docs/plugins/cli-backend-plugins.md

@@ -197,22 +197,30 @@ only for behavior that really belongs to the backend.
 
 `CliBackendPlugin` can also define:
 
-| Hook                               | Use                                                    |
-| ---------------------------------- | ------------------------------------------------------ |
-| `normalizeConfig(config, context)` | Rewrite legacy user config after merge                 |
-| `resolveExecutionArgs(ctx)`        | Add request-scoped flags such as thinking effort       |
-| `prepareExecution(ctx)`            | Create temporary auth or config bridges before launch  |
-| `transformSystemPrompt(ctx)`       | Apply a final CLI-specific system prompt transform     |
-| `textTransforms`                   | Bidirectional prompt/output replacements               |
-| `defaultAuthProfileId`             | Prefer a specific OpenClaw auth profile                |
-| `authEpochMode`                    | Decide how auth changes invalidate stored CLI sessions |
-| `nativeToolMode`                   | Declare whether the CLI has always-on native tools     |
-| `bundleMcp` / `bundleMcpMode`      | Opt into OpenClaw's loopback MCP tool bridge           |
-| `ownsNativeCompaction`             | Backend owns its own compaction - OpenClaw defers      |
+| Hook                               | Use                                                                         |
+| ---------------------------------- | --------------------------------------------------------------------------- |
+| `normalizeConfig(config, context)` | Rewrite legacy user config after merge                                      |
+| `resolveExecutionArgs(ctx)`        | Add request-scoped flags such as thinking effort or side-question isolation |
+| `prepareExecution(ctx)`            | Create temporary auth or config bridges before launch                       |
+| `transformSystemPrompt(ctx)`       | Apply a final CLI-specific system prompt transform                          |
+| `textTransforms`                   | Bidirectional prompt/output replacements                                    |
+| `defaultAuthProfileId`             | Prefer a specific OpenClaw auth profile                                     |
+| `authEpochMode`                    | Decide how auth changes invalidate stored CLI sessions                      |
+| `nativeToolMode`                   | Declare whether the CLI has always-on native tools                          |
+| `sideQuestionToolMode`             | Declare disabled native tools for `/btw` side questions                     |
+| `bundleMcp` / `bundleMcpMode`      | Opt into OpenClaw's loopback MCP tool bridge                                |
+| `ownsNativeCompaction`             | Backend owns its own compaction - OpenClaw defers                           |
 
 Keep these hooks provider-owned. Do not add CLI-specific branches to core when a
 backend hook can express the behavior.
 
+`ctx.executionMode` is `"agent"` for normal turns and `"side-question"` for
+ephemeral `/btw` calls. Use it when the CLI needs different one-shot flags, such
+as disabling native tools, session persistence, or resume behavior for BTW. If a
+backend normally has `nativeToolMode: "always-on"` but its side-question argv
+reliably disables those tools, also set `sideQuestionToolMode: "disabled"`;
+otherwise OpenClaw fails closed when BTW requires a no-tools CLI run.
+
 ### `ownsNativeCompaction`: opting out of OpenClaw compaction
 
 If your backend runs an agent that compacts its **own** transcript, set

docs/plugins/codex-harness-reference.md

@@ -313,9 +313,13 @@ available timeout in this order:
 - For `image_generate` without a configured timeout, the 120 second
   image-generation default.
 - For the media-understanding `image` tool, `tools.media.image.timeoutSeconds`
-  converted to milliseconds, or the 60 second media default.
+  converted to milliseconds, or the 60 second media default. For image
+  understanding, this applies to the request itself and is not reduced by
+  earlier preparation work.
 - The 90 second dynamic-tool default.
 
+This watchdog is the outer dynamic `item/tool/call` budget. Provider-specific
+request timeouts run inside that call and keep their own timeout semantics.
 Dynamic tool budgets are capped at 600000 ms. On timeout, OpenClaw aborts the
 tool signal where supported and returns a failed dynamic-tool response to Codex
 so the turn can continue instead of leaving the session in `processing`.

docs/plugins/codex-harness.md

@@ -557,10 +557,14 @@ or shortens that specific tool budget. The `image_generate` tool uses
 `agents.defaults.imageGenerationModel.timeoutMs` when the tool call does not
 provide its own timeout, or a 120 second image-generation default otherwise.
 The media-understanding `image` tool uses
-`tools.media.image.timeoutSeconds` or its 60 second media default. Dynamic tool
-budgets are capped at 600000 ms. On timeout, OpenClaw aborts the tool signal
+`tools.media.image.timeoutSeconds` or its 60 second media default. For image
+understanding, that timeout applies to the request itself and is not
+reduced by earlier preparation work. Dynamic tool budgets are
+capped at 600000 ms. On timeout, OpenClaw aborts the tool signal
 where supported and returns a failed dynamic-tool response to Codex so the turn
 can continue instead of leaving the session in `processing`.
+This watchdog is the outer dynamic `item/tool/call` budget; provider-specific
+request timeouts run inside that call and keep their own timeout semantics.
 
 After Codex accepts a turn, and after OpenClaw responds to a turn-scoped
 app-server request, the harness expects Codex to make current-turn progress and

docs/plugins/hooks.md

@@ -152,7 +152,8 @@ observation-only.
 - `gateway_start` / `gateway_stop` - start or stop plugin-owned services with the Gateway
 - `deactivate` - deprecated compatibility alias for `gateway_stop`; use `gateway_stop` in new plugins
 - `cron_changed` - observe gateway-owned cron lifecycle changes (added, updated, removed, started, finished, scheduled)
-- **`before_install`** - inspect skill or plugin install context and optionally block
+- **`before_install`** - inspect staged skill or plugin install material from a loaded
+  plugin runtime
 
 ## Debug runtime hooks
 
@@ -462,11 +463,19 @@ Decision rules:
 
 ## Install hooks
 
-`before_install` runs after the operator-owned `security.installPolicy` check
-when one is configured. The `builtinScan` field remains in the event payload for
-compatibility, but OpenClaw no longer runs built-in install-time dangerous-code
-blocking, so it is an empty `ok` result. Return additional findings or
-`{ block: true, blockReason }` to stop the install.
+Use `security.installPolicy` for operator-owned allow/block decisions. That
+policy runs from OpenClaw config, covers CLI install and update paths, and fails
+closed when enabled but unavailable.
+
+`before_install` is a plugin-runtime lifecycle hook. It runs after
+`security.installPolicy` only in the OpenClaw process where plugin hooks have
+already been loaded, such as Gateway-backed install flows. It is useful for
+plugin-owned observations, warnings, and compatibility checks, but it is not the
+primary enterprise or host security boundary for installs. The `builtinScan`
+field remains in the event payload for compatibility, but OpenClaw no longer
+runs built-in install-time dangerous-code blocking, so it is an empty `ok`
+result. Return additional findings or `{ block: true, blockReason }` to stop the
+install in that process.
 
 `block: true` is terminal. `block: false` is treated as no decision.
 Handler failures block the install fail-closed.

docs/plugins/manifest.md

@@ -1278,6 +1278,7 @@ Important examples:
 | `openclaw.compat.pluginApi`                                                                | Minimum OpenClaw plugin API range required by this package, using a semver floor like `>=2026.5.27`.                                                                                 |
 | `openclaw.install.expectedIntegrity`                                                       | Expected npm dist integrity string such as `sha512-...`; install and update flows verify the fetched artifact against it.                                                            |
 | `openclaw.install.allowInvalidConfigRecovery`                                              | Allows a narrow bundled-plugin reinstall recovery path when config is invalid.                                                                                                       |
+| `openclaw.install.requiredPlatformPackages`                                                | npm package aliases that must materialize when their lockfile platform constraints match the current host.                                                                           |
 | `openclaw.startup.deferConfiguredChannelFullLoadUntilAfterListen`                          | Lets setup-runtime channel surfaces load before listen, then defers the full configured channel plugin until post-listen activation.                                                 |
 
 Manifest metadata decides which provider/channel/setup choices appear in
@@ -1290,6 +1291,13 @@ registry loading for non-bundled plugin sources. Invalid values are rejected;
 newer-but-valid values skip external plugins on older hosts. Bundled source
 plugins are assumed to be co-versioned with the host checkout.
 
+`openclaw.install.requiredPlatformPackages` is for npm packages that expose
+required native binaries through optional, platform-specific aliases. List the
+bare npm package name for every supported platform alias. During npm install,
+OpenClaw verifies only the declared alias whose lockfile constraints match the
+current host. If npm reports success but omits that alias, OpenClaw retries once
+with a fresh cache and rolls back the install if the alias is still missing.
+
 `openclaw.compat.pluginApi` is enforced during package install for non-bundled
 plugin sources. Use it for the OpenClaw plugin SDK/runtime API floor that the
 package was built against. It can be stricter than `minHostVersion` when a

docs/plugins/sdk-overview.md

@@ -378,7 +378,10 @@ AI CLI backend such as `claude-cli` or `my-cli`.
   (for example normalizing old flag shapes).
 - Use `resolveExecutionArgs` for request-scoped argv rewrites that belong to
   the CLI dialect, such as mapping OpenClaw thinking levels to a native effort
-  flag.
+  flag. The hook receives `ctx.executionMode`; use `"side-question"` to add
+  backend-native isolation flags for ephemeral `/btw` calls. If those flags
+  reliably disable native tools for an otherwise always-on CLI, declare
+  `sideQuestionToolMode: "disabled"` too.
 
 For an end-to-end authoring guide, see
 [CLI backend plugins](/plugins/cli-backend-plugins).
@@ -428,6 +431,10 @@ semantics.
 
 ### Hook decision semantics
 
+`before_install` is a plugin-runtime lifecycle hook, not the operator install
+policy surface. Use `security.installPolicy` when an allow/block decision must
+cover CLI and Gateway-backed install or update paths.
+
 - `before_tool_call`: returning `{ block: true }` is terminal. Once any handler sets it, lower-priority handlers are skipped.
 - `before_tool_call`: returning `{ block: false }` is treated as no decision (same as omitting `block`), not as an override.
 - `before_install`: returning `{ block: true }` is terminal. Once any handler sets it, lower-priority handlers are skipped.

docs/plugins/sdk-provider-plugins.md

@@ -515,6 +515,7 @@ API key auth, and dynamic model resolution.
 
       - `openclaw/plugin-sdk/provider-model-shared` - `ProviderReplayFamily`, `buildProviderReplayFamilyHooks(...)`, and the raw replay builders (`buildOpenAICompatibleReplayPolicy`, `buildAnthropicReplayPolicyForModel`, `buildGoogleGeminiReplayPolicy`, `buildHybridAnthropicOrOpenAIReplayPolicy`). Also exports Gemini replay helpers (`sanitizeGoogleGeminiReplayHistory`, `resolveTaggedReasoningOutputMode`) and endpoint/model helpers (`resolveProviderEndpoint`, `normalizeProviderId`, `normalizeGooglePreviewModelId`).
       - `openclaw/plugin-sdk/provider-stream` - `ProviderStreamFamily`, `buildProviderStreamFamilyHooks(...)`, `composeProviderStreamWrappers(...)`, plus the shared OpenAI/Codex wrappers (`createOpenAIAttributionHeadersWrapper`, `createOpenAIFastModeWrapper`, `createOpenAIServiceTierWrapper`, `createOpenAIResponsesContextManagementWrapper`, `createCodexNativeWebSearchWrapper`), DeepSeek V4 OpenAI-compatible wrapper (`createDeepSeekV4OpenAICompatibleThinkingWrapper`), Anthropic Messages thinking prefill cleanup (`createAnthropicThinkingPrefillPayloadWrapper`), plain-text tool-call compat (`createPlainTextToolCallCompatWrapper`), and shared proxy/provider wrappers (`createOpenRouterWrapper`, `createToolStreamWrapper`, `createMinimaxFastModeWrapper`).
+      - `openclaw/plugin-sdk/provider-stream-shared` - lightweight payload and event wrappers for hot provider paths, including `createOpenAICompatibleCompletionsThinkingOffWrapper`, `createPayloadPatchStreamWrapper`, and `createPlainTextToolCallCompatWrapper`.
       - `openclaw/plugin-sdk/provider-tools` - `ProviderToolCompatFamily`, `buildProviderToolCompatFamilyHooks("deepseek" | "gemini" | "openai")`, and underlying provider schema helpers.
 
       For Gemini-family providers, keep the reasoning-output mode aligned with

docs/plugins/sdk-setup.md

@@ -163,6 +163,7 @@ Example:
 | `minHostVersion`             | `string`                            | Minimum supported OpenClaw version in the form `>=x.y.z` or `>=x.y.z-prerelease`. |
 | `expectedIntegrity`          | `string`                            | Expected npm dist integrity string, usually `sha512-...`, for pinned installs.    |
 | `allowInvalidConfigRecovery` | `boolean`                           | Lets bundled-plugin reinstall flows recover from specific stale-config failures.  |
+| `requiredPlatformPackages`   | `string[]`                          | Required platform-specific npm aliases verified during npm install.               |
 
 <AccordionGroup>
   <Accordion title="Onboarding behavior">

docs/plugins/sdk-subpaths.md

@@ -164,7 +164,7 @@ and pairing-path families.
     | `plugin-sdk/provider-tools` | `ProviderToolCompatFamily`, `buildProviderToolCompatFamilyHooks`, and DeepSeek/Gemini/OpenAI schema cleanup + diagnostics |
     | `plugin-sdk/provider-usage` | Provider usage snapshot types, shared usage fetch helpers, and provider fetchers such as `fetchClaudeUsage` |
     | `plugin-sdk/provider-stream` | `ProviderStreamFamily`, `buildProviderStreamFamilyHooks`, `composeProviderStreamWrappers`, stream wrapper types, plain-text tool-call compat, and shared Anthropic/Bedrock/DeepSeek V4/Google/Kilocode/Moonshot/OpenAI/OpenRouter/Z.A.I/MiniMax/Copilot wrapper helpers |
-    | `plugin-sdk/provider-stream-shared` | Public shared provider stream wrapper helpers including `composeProviderStreamWrappers`, `createPlainTextToolCallCompatWrapper`, `createPayloadPatchStreamWrapper`, `createToolStreamWrapper`, and Anthropic/DeepSeek/OpenAI-compatible stream utilities |
+    | `plugin-sdk/provider-stream-shared` | Public shared provider stream wrapper helpers including `composeProviderStreamWrappers`, `createOpenAICompatibleCompletionsThinkingOffWrapper`, `createPlainTextToolCallCompatWrapper`, `createPayloadPatchStreamWrapper`, `createToolStreamWrapper`, and Anthropic/DeepSeek/OpenAI-compatible stream utilities |
     | `plugin-sdk/provider-transport-runtime` | Native provider transport helpers such as guarded fetch, transport message transforms, and writable transport event streams |
     | `plugin-sdk/provider-onboard` | Onboarding config patch helpers |
     | `plugin-sdk/global-singleton` | Process-local singleton/map/cache helpers |
@@ -236,6 +236,7 @@ usage endpoint failed or returned no usable usage data.
     | `plugin-sdk/config-contracts` | Focused type-only config surface for plugin config shapes such as `OpenClawConfig` and channel/provider config types |
     | `plugin-sdk/plugin-config-runtime` | Runtime plugin-config lookup helpers such as `requireRuntimeConfig`, `resolvePluginConfigObject`, and `resolveLivePluginConfigObject` |
     | `plugin-sdk/config-mutation` | Transactional config mutation helpers such as `mutateConfigFile`, `replaceConfigFile`, and `logConfigUpdated` |
+    | `plugin-sdk/message-tool-delivery-hints` | Shared message-tool delivery metadata hint strings |
     | `plugin-sdk/runtime-config-snapshot` | Current process config snapshot helpers such as `getRuntimeConfig`, `getRuntimeConfigSnapshot`, and test snapshot setters |
     | `plugin-sdk/telegram-command-config` | Telegram command-name/description normalization and duplicate/conflict checks, even when the bundled Telegram contract surface is unavailable |
     | `plugin-sdk/text-autolink-runtime` | File-reference autolink detection without the broad text barrel |

docs/providers/zai.md

@@ -19,7 +19,7 @@ OpenClaw uses the `zai` provider with a Z.AI API key.
 ## GLM models
 
 GLM is a model family, not a separate provider. In OpenClaw, GLM models use
-refs such as `zai/glm-5.1`: provider `zai`, model id `glm-5.1`.
+refs such as `zai/glm-5.2`: provider `zai`, model id `glm-5.2`.
 
 ## Getting started
 
@@ -85,12 +85,12 @@ you want to force a specific Coding Plan or general API surface.
   models: {
     providers: {
       zai: {
-        // Example value. Onboarding writes the matching baseUrl for your endpoint.
-        baseUrl: "https://api.z.ai/api/paas/v4",
+        // GLM-5.2 uses the Coding Plan endpoint.
+        baseUrl: "https://api.z.ai/api/coding/paas/v4",
       },
     },
   },
-  agents: { defaults: { model: { primary: "zai/glm-5.1" } } },
+  agents: { defaults: { model: { primary: "zai/glm-5.2" } } },
 }
 ```
 
@@ -105,28 +105,31 @@ openclaw models list --all --provider zai
 
 The manifest-backed catalog currently includes:
 
-| Model ref            | Notes         |
-| -------------------- | ------------- |
-| `zai/glm-5.1`        | Default model |
-| `zai/glm-5`          |               |
-| `zai/glm-5-turbo`    |               |
-| `zai/glm-5v-turbo`   |               |
-| `zai/glm-4.7`        |               |
-| `zai/glm-4.7-flash`  |               |
-| `zai/glm-4.7-flashx` |               |
-| `zai/glm-4.6`        |               |
-| `zai/glm-4.6v`       |               |
-| `zai/glm-4.5`        |               |
-| `zai/glm-4.5-air`    |               |
-| `zai/glm-4.5-flash`  |               |
-| `zai/glm-4.5v`       |               |
+| Model ref            | Notes                           |
+| -------------------- | ------------------------------- |
+| `zai/glm-5.2`        | Coding Plan default; 1M context |
+| `zai/glm-5.1`        | General API default             |
+| `zai/glm-5`          |                                 |
+| `zai/glm-5-turbo`    |                                 |
+| `zai/glm-5v-turbo`   |                                 |
+| `zai/glm-4.7`        |                                 |
+| `zai/glm-4.7-flash`  |                                 |
+| `zai/glm-4.7-flashx` |                                 |
+| `zai/glm-4.6`        |                                 |
+| `zai/glm-4.6v`       |                                 |
+| `zai/glm-4.5`        |                                 |
+| `zai/glm-4.5-air`    |                                 |
+| `zai/glm-4.5-flash`  |                                 |
+| `zai/glm-4.5v`       |                                 |
 
 <Tip>
 GLM models are available as `zai/<model>` (example: `zai/glm-5`).
 </Tip>
 
 <Note>
-The default bundled model ref is `zai/glm-5.1`. GLM versions and availability
+Coding Plan setup defaults to `zai/glm-5.2`; general API setup keeps
+`zai/glm-5.1`. Endpoint auto-detection falls back to `glm-5.1` or `glm-4.7`
+when the selected plan does not expose GLM-5.2. GLM versions and availability
 can change; run `openclaw models list --all --provider zai` to see the catalog
 known to your installed version.
 </Note>
@@ -173,7 +176,7 @@ known to your installed version.
       agents: {
         defaults: {
           models: {
-            "zai/glm-5.1": {
+            "zai/glm-5.2": {
               params: { preserveThinking: true },
             },
           },

docs/reference/RELEASING.md

@@ -99,10 +99,14 @@ the maintainer-only release runbook.
    file, lane, workflow job, package profile, provider, or model allowlist that
    proves the fix. Rerun the full umbrella only when the changed surface makes
    prior evidence stale.
-9. For beta, tag `vYYYY.M.PATCH-beta.N`, then run `pnpm release:candidate -- --tag
-vYYYY.M.PATCH-beta.N` from the matching `release/YYYY.M.PATCH` branch. The helper runs
-   the local generated-release checks, dispatches or verifies the full release
-   validation and npm preflight evidence, runs Parallels and Telegram package
+9. For a tagged beta candidate, run
+   `pnpm release:candidate -- --tag vYYYY.M.PATCH-beta.N` from the matching
+   `release/YYYY.M.PATCH` branch. For stable, pass the required Windows source
+   release too:
+   `pnpm release:candidate -- --tag vYYYY.M.PATCH --windows-node-tag vX.Y.Z`.
+   The helper runs the local generated-release checks, dispatches or verifies
+   the full release validation and npm preflight evidence, runs Parallels
+   fresh/update proof against the exact prepared tarball plus Telegram package
    proof, records plugin npm and ClawHub plans, and prints the exact
    `OpenClaw Release Publish` command only after the evidence bundle is green.
    `OpenClaw Release Publish` dispatches the selected or all-publishable plugin
@@ -142,9 +146,12 @@ vYYYY.M.PATCH-beta.N` from the matching `release/YYYY.M.PATCH` branch. The helpe
     direct push, it opens or updates an appcast PR. Stable Windows Hub
     readiness requires the signed `OpenClawCompanion-Setup-x64.exe`,
     `OpenClawCompanion-Setup-arm64.exe`, and
-    `OpenClawCompanion-SHA256SUMS.txt` assets on the OpenClaw GitHub release;
-    promote them with the `Windows Node Release` workflow after the matching
-    `openclaw/openclaw-windows-node` release has passed its signing workflow.
+    `OpenClawCompanion-SHA256SUMS.txt` assets on the OpenClaw GitHub release.
+    Pass the exact signed `openclaw/openclaw-windows-node` release tag as
+    `windows_node_tag` and its candidate-approved installer digest map as
+    `windows_node_installer_digests`; `OpenClaw Release Publish` keeps the
+    release draft, dispatches `Windows Node Release`, and verifies all three
+    assets before publication.
 11. After publish, run the npm post-publish verifier, optional standalone
     published-npm Telegram E2E when you need post-publish channel proof,
     dist-tag promotion when needed, verify the generated GitHub release page,
@@ -253,21 +260,36 @@ vYYYY.M.PATCH-beta.N` from the matching `release/YYYY.M.PATCH` branch. The helpe
   to the GitHub release as `openclaw-<version>-dependency-evidence.zip`.
 - Run `OpenClaw Release Publish` for the mutating publish sequence after the
   tag exists. Dispatch it from `release/YYYY.M.PATCH` (or `main` when publishing a
-  main-reachable tag), pass the release tag and successful OpenClaw npm
-  `preflight_run_id`, and keep the default plugin publish scope
-  `all-publishable` unless you are deliberately running a focused repair. The
-  workflow serializes plugin npm publish, plugin ClawHub publish, and OpenClaw
-  npm publish so the core package is not published before its externalized
-  plugins.
-- Run the manual `Windows Node Release` workflow for stable releases after the
-  matching `openclaw/openclaw-windows-node` release exists. It downloads the
-  signed Windows Hub installers from the companion repo, verifies their
-  Authenticode signatures on a Windows runner, writes a SHA-256 manifest, and
-  uploads the installers plus manifest onto the canonical OpenClaw GitHub
-  release. Website download links should target exact OpenClaw release asset
-  URLs for the current stable release, or `releases/latest/download/...` only
-  after verifying GitHub's latest redirect points at that same release; do not
-  link only to the companion repo release page.
+  main-reachable tag), pass the release tag, successful OpenClaw npm
+  `preflight_run_id`, and successful `full_release_validation_run_id`, and keep
+  the default plugin publish scope `all-publishable` unless you are deliberately
+  running a focused repair. The workflow serializes plugin npm publish, plugin
+  ClawHub publish, and OpenClaw npm publish so the core package is not published
+  before its externalized plugins.
+- Stable `OpenClaw Release Publish` requires an exact `windows_node_tag` after
+  the matching non-prerelease `openclaw/openclaw-windows-node` release exists.
+  It also requires the candidate-approved `windows_node_installer_digests` map.
+  Before dispatching any publish child, it verifies that source release is
+  published, non-prerelease, contains the required x64/ARM64 installers, and
+  still matches that approved map. It then dispatches `Windows Node Release`
+  while the OpenClaw release is still a draft, carrying the pinned installer
+  digest map unchanged. The child
+  workflow downloads the signed Windows Hub installers from that exact tag,
+  matches them against the pinned digests, verifies their Authenticode
+  signatures use the expected OpenClaw Foundation signer on a Windows runner,
+  writes a SHA-256 manifest, and uploads the installers plus manifest onto the
+  canonical OpenClaw GitHub release, then re-downloads the promoted assets and
+  verifies the manifest membership and hashes. The parent verifies the current
+  x64, ARM64, and checksum asset contract before publication. Direct recovery
+  rejects unexpected `OpenClawCompanion-*` asset names before replacing the
+  expected contract assets with the pinned source bytes. Manually dispatch
+  `Windows Node Release` only for recovery, and always pass an exact tag, never
+  `latest`, plus the explicit `expected_installer_digests` JSON map from the
+  approved source release. Website download links should target exact OpenClaw
+  release asset URLs for the current stable release, or
+  `releases/latest/download/...` only after verifying GitHub's latest redirect
+  points at that same release; do not link only to the companion repo release
+  page.
 - Release checks now run in a separate manual workflow:
   `OpenClaw Release Checks`
 - `OpenClaw Release Checks` also runs the QA Lab mock parity lane plus the fast
@@ -697,7 +719,12 @@ orchestrates the trusted-publisher workflows in the order the release needs:
    `ref=<release-sha>`.
 5. Dispatch `Plugin ClawHub Release` with the same scope and SHA.
 6. Dispatch `OpenClaw NPM Release` with the release tag, npm dist-tag, and
-   saved `preflight_run_id`.
+   saved `preflight_run_id` after verifying the saved
+   `full_release_validation_run_id`.
+7. For stable releases, create or update the GitHub release as a draft, dispatch
+   `Windows Node Release` with the explicit `windows_node_tag` and
+   candidate-approved `windows_node_installer_digests`, and verify the canonical
+   installer/checksum assets before publishing the draft.
 
 Beta publish example:
 
@@ -706,6 +733,7 @@ gh workflow run openclaw-release-publish.yml \
   --ref release/YYYY.M.PATCH \
   -f tag=vYYYY.M.PATCH-beta.N \
   -f preflight_run_id=<successful-openclaw-npm-preflight-run-id> \
+  -f full_release_validation_run_id=<successful-full-release-validation-run-id> \
   -f npm_dist_tag=beta
 ```
 
@@ -715,7 +743,10 @@ Stable publish to the default beta dist-tag:
 gh workflow run openclaw-release-publish.yml \
   --ref release/YYYY.M.PATCH \
   -f tag=vYYYY.M.PATCH \
+  -f windows_node_tag=vX.Y.Z \
+  -f windows_node_installer_digests='{"OpenClawCompanion-Setup-x64.exe":"sha256:<approved-x64-sha256>","OpenClawCompanion-Setup-arm64.exe":"sha256:<approved-arm64-sha256>"}' \
   -f preflight_run_id=<successful-openclaw-npm-preflight-run-id> \
+  -f full_release_validation_run_id=<successful-full-release-validation-run-id> \
   -f npm_dist_tag=beta
 ```
 
@@ -725,7 +756,10 @@ Stable promotion directly to `latest` is explicit:
 gh workflow run openclaw-release-publish.yml \
   --ref release/YYYY.M.PATCH \
   -f tag=vYYYY.M.PATCH \
+  -f windows_node_tag=vX.Y.Z \
+  -f windows_node_installer_digests='{"OpenClawCompanion-Setup-x64.exe":"sha256:<approved-x64-sha256>","OpenClawCompanion-Setup-arm64.exe":"sha256:<approved-arm64-sha256>"}' \
   -f preflight_run_id=<successful-openclaw-npm-preflight-run-id> \
+  -f full_release_validation_run_id=<successful-full-release-validation-run-id> \
   -f npm_dist_tag=latest
 ```
 
@@ -755,6 +789,13 @@ package cannot ship without every publishable official plugin, including
 - `tag`: required release tag; must already exist
 - `preflight_run_id`: successful `OpenClaw NPM Release` preflight run id;
   required when `publish_openclaw_npm=true`
+- `full_release_validation_run_id`: successful `Full Release Validation` run
+  id; required when `publish_openclaw_npm=true`
+- `windows_node_tag`: exact non-prerelease `openclaw/openclaw-windows-node`
+  release tag; required for stable OpenClaw publish
+- `windows_node_installer_digests`: candidate-approved compact JSON map of the
+  current Windows installer names to their pinned `sha256:` digests; required
+  for stable OpenClaw publish
 - `npm_dist_tag`: npm target tag for the OpenClaw package
 - `plugin_publish_scope`: defaults to `all-publishable`; use `selected` only
   for focused plugin-only repair work with `publish_openclaw_npm=false`
@@ -800,14 +841,21 @@ When cutting a stable npm release:
    Matrix, and Telegram coverage from one manual workflow
 4. If you intentionally only need the deterministic normal test graph, run the
    manual `CI` workflow on the release ref instead
-5. Save the successful `preflight_run_id`
-6. Run `OpenClaw Release Publish` with the same `tag`, the same `npm_dist_tag`,
-   and the saved `preflight_run_id`; it publishes externalized plugins to npm
-   and ClawHub before promoting the OpenClaw npm package
-7. If the release landed on `beta`, use the
+5. Select the exact non-prerelease `openclaw/openclaw-windows-node` release tag
+   whose signed x64 and ARM64 installers should ship. Save it as
+   `windows_node_tag`, and save their validated digest map as
+   `windows_node_installer_digests`. The release-candidate helper records both
+   and includes them in its generated publish command.
+6. Save the successful `preflight_run_id` and `full_release_validation_run_id`
+7. Run `OpenClaw Release Publish` with the same `tag`, the same `npm_dist_tag`,
+   the selected `windows_node_tag`, its saved `windows_node_installer_digests`,
+   the saved `preflight_run_id`, and the saved `full_release_validation_run_id`;
+   it publishes externalized plugins to npm and ClawHub before promoting the
+   OpenClaw npm package
+8. If the release landed on `beta`, use the
    `openclaw/releases/.github/workflows/openclaw-npm-dist-tags.yml`
    workflow to promote that stable version from `beta` to `latest`
-8. If the release intentionally published directly to `latest` and `beta`
+9. If the release intentionally published directly to `latest` and `beta`
    should follow the same stable build immediately, use that same release
    workflow to point both dist-tags at the stable version, or let its scheduled
    self-healing sync move `beta` later

docs/reference/api-usage-costs.md

@@ -140,7 +140,7 @@ See [Memory](/concepts/memory).
 - **Ollama Web Search**: key-free for a reachable signed-in local Ollama host; direct `https://ollama.com` search uses `OLLAMA_API_KEY`, and auth-protected hosts can reuse normal Ollama provider bearer auth
 - **Perplexity Search API**: `PERPLEXITY_API_KEY`, `OPENROUTER_API_KEY`, or `plugins.entries.perplexity.config.webSearch.apiKey`
 - **Tavily**: `TAVILY_API_KEY` or `plugins.entries.tavily.config.webSearch.apiKey`
-- **DuckDuckGo**: key-free fallback (no API billing, but unofficial and HTML-based)
+- **DuckDuckGo**: key-free provider when explicitly selected (no API billing, but unofficial and HTML-based)
 - **SearXNG**: `SEARXNG_BASE_URL` or `plugins.entries.searxng.config.webSearch.baseUrl` (key-free/self-hosted; no hosted API billing)
 
 Legacy `tools.web.search.*` provider paths still load through the temporary compatibility shim, but they are no longer the recommended config surface.

docs/tools/acp-agents.md

@@ -336,6 +336,7 @@ top-level `bindings[]` entries.
 - **Discord channel/thread:** `match.channel="discord"` + `match.peer.id="<channelOrThreadId>"`
 - **Slack channel/DM:** `match.channel="slack"` + `match.peer.id="<channelId|channel:<channelId>|#<channelId>|userId|user:<userId>|slack:<userId>|<@userId>>"`. Prefer stable Slack ids; channel bindings also match replies inside that channel's threads.
 - **Telegram forum topic:** `match.channel="telegram"` + `match.peer.id="<chatId>:topic:<topicId>"`
+- **WhatsApp DM/group:** `match.channel="whatsapp"` + `match.peer.id="<E.164|group JID>"`. Use E.164 numbers such as `+15555550123` for direct chats and WhatsApp group JIDs such as `120363424282127706@g.us` for groups.
 - **iMessage DM/group:** `match.channel="imessage"` + `match.peer.id="<handle|chat_id:*|chat_guid:*|chat_identifier:*>"`. Prefer `chat_id:*` for stable group bindings.
 
 </ParamField>
@@ -453,8 +454,9 @@ Use `agents.list[].runtime` to define ACP defaults once per agent:
 
 ### Behavior
 
-- OpenClaw ensures the configured ACP session exists before use.
-- Messages in that channel or topic route to the configured ACP session.
+- OpenClaw ensures the configured ACP session exists after channel-specific admission and before use.
+- Messages in that channel, topic, or chat route to the configured ACP session.
+- Configured ACP bindings own their session route. Channel broadcast fan-out does not replace the configured ACP session for a matched binding.
 - In bound conversations, `/new` and `/reset` reset the same ACP session key in place.
 - Temporary runtime bindings (for example created by thread-focus flows) still apply where present.
 - For cross-agent ACP spawns without an explicit `cwd`, OpenClaw inherits the target agent workspace from agent config.

docs/tools/browser-control.md

@@ -258,7 +258,14 @@ Snapshot flags at a glance:
 - `--format aria`: accessibility tree with `axN` refs. When Playwright is available, OpenClaw binds refs with backend DOM ids to the live page so follow-up actions can use them; otherwise treat the output as inspection-only.
 - `--efficient` (or `--mode efficient`): compact role snapshot preset. Set `browser.snapshotDefaults.mode: "efficient"` to make this the default (see [Gateway configuration](/gateway/configuration-reference#browser)).
 - `--interactive`, `--compact`, `--depth`, `--selector` force a role snapshot with `ref=e12` refs. `--frame "<iframe>"` scopes role snapshots to an iframe.
-- `--labels` adds a viewport-only screenshot with overlayed ref labels and prints the saved path.
+- With Playwright, `--labels` adds a screenshot with overlayed ref labels
+  (prints `MEDIA:<path>`) plus an `annotations` array with each ref's bounding
+  box. On `screenshot`, Playwright-backed labels work with `--full-page`,
+  `--ref`, and `--element`; on `snapshot`, the accompanying screenshot remains
+  viewport-only. Existing-session/chrome-mcp profiles render overlay labels on
+  page screenshots but do not return `annotations` or use the Playwright
+  full-page/ref/element projection helper. Without Playwright or chrome-mcp,
+  labeled screenshots are not available.
 - `--urls` appends discovered link destinations to AI snapshots.
 
 ## Snapshots and refs
@@ -274,7 +281,9 @@ OpenClaw supports two "snapshot" styles:
   - Output: a role-based list/tree with `[ref=e12]` (and optional `[nth=1]`).
   - Actions: `openclaw browser click e12`, `openclaw browser highlight e12`.
   - Internally, the ref is resolved via `getByRole(...)` (plus `nth()` for duplicates).
-  - Add `--labels` to include a viewport screenshot with overlayed `e12` labels.
+  - Add `--labels` to include a screenshot with overlayed `e12` labels. On
+    Playwright-backed profiles this also returns per-ref bounding-box metadata
+    (`annotations[]`).
   - Add `--urls` when link text is ambiguous and the agent needs concrete
     navigation targets.
 

docs/tools/btw.md

@@ -42,8 +42,14 @@ app-server thread as an ephemeral side thread. That keeps Codex OAuth and native
 thread behavior intact while still isolating the side answer from the parent
 transcript. Like Codex `/side`, the side thread keeps the current Codex
 permissions and native tool surface, with guardrails that tell the model not to
-treat inherited parent-thread work as active instructions. Non-Codex runtimes
-keep the older direct one-shot path.
+treat inherited parent-thread work as active instructions.
+
+For CLI runtime aliases, BTW uses the owning CLI backend in side-question mode
+instead of falling back to a direct provider call. OpenClaw seeds sanitized
+conversation context into a fresh one-shot CLI invocation, disables OpenClaw MCP
+tool bundling and reusable CLI session state for that invocation, and lets the
+backend add any CLI-native no-resume or no-tools flags it supports. Direct
+non-CLI runtimes keep the direct one-shot path.
 
 ## What it does not do
 

docs/tools/duckduckgo-search.md

@@ -1,9 +1,9 @@
 ---
-summary: "DuckDuckGo web search -- key-free fallback provider (experimental, HTML-based)"
+summary: "DuckDuckGo web search -- key-free provider (experimental, HTML-based)"
 read_when:
   - You want a web search provider that requires no API key
   - You want to use DuckDuckGo for web_search
-  - You need a zero-config search fallback
+  - You want an explicitly selected key-free search provider
 title: "DuckDuckGo search"
 ---
 
@@ -85,16 +85,16 @@ parameters override config values per-query.
 
 ## Notes
 
-- **No API key** - works out of the box, zero configuration
+- **No API key** - works after you select DuckDuckGo as your `web_search`
+  provider
 - **Experimental** - gathers results from DuckDuckGo's non-JavaScript HTML
   search pages, not an official API or SDK
 - **Bot-challenge risk** - DuckDuckGo may serve CAPTCHAs or block requests
   under heavy or automated use
 - **HTML parsing** - results depend on page structure, which can change without
   notice
-- **Auto-detection order** - DuckDuckGo is the first key-free fallback
-  (order 100) in auto-detection. API-backed providers with configured keys run
-  first, then Ollama Web Search (order 110), then SearXNG (order 200)
+- **Explicit selection** - OpenClaw does not choose DuckDuckGo automatically
+  when no API-backed provider is configured
 - **SafeSearch defaults to moderate** when not configured
 
 <Tip>

docs/tools/ollama-search.md

@@ -140,8 +140,9 @@ Direct hosted Ollama Web Search:
   that env key to the local host.
 - OpenClaw warns during setup if Ollama is unreachable or not signed in, but
   it does not block selection.
-- Runtime auto-detect can fall back to Ollama Web Search when no higher-priority
-  credentialed provider is configured.
+- OpenClaw does not auto-select Ollama Web Search when no higher-priority
+  credentialed provider is configured; choose it explicitly with
+  `tools.web.search.provider: "ollama"`.
 - Local Ollama daemon hosts use the local proxy endpoint
   `/api/experimental/web_search`, which signs and forwards to Ollama Cloud.
 - `https://ollama.com` hosts use the public hosted endpoint

docs/tools/parallel-search.md

@@ -11,8 +11,8 @@ OpenClaw bundles two [Parallel](https://parallel.ai/) `web_search` providers:
 
 - **Parallel Search (Free)** (`parallel-free`) -- Parallel's free
   [Search MCP](https://docs.parallel.ai/integrations/mcp/search-mcp). Requires no
-  account or API key. OpenClaw selects it automatically when no other web search
-  provider is configured, so `web_search` works without setup.
+  account or API key. Select it explicitly when you want Parallel's hosted
+  key-free search path.
 - **Parallel Search** (`parallel`) -- Parallel's paid Search API. Requires a
   `PARALLEL_API_KEY` and offers higher rate limits and objective tuning.
 
@@ -29,8 +29,8 @@ explicitly.
 
 ## API key (paid provider)
 
-`parallel-free` requires no setup. The paid `parallel` provider needs an API
-key:
+`parallel-free` requires no API key, but it still must be selected as the
+managed provider. The paid `parallel` provider needs an API key:
 
 <Steps>
   <Step title="Create an account">
@@ -66,6 +66,8 @@ key:
   tools: {
     web: {
       search: {
+        // Use "parallel-free" for the free Search MCP, or "parallel" for
+        // the paid API-backed provider shown here.
         provider: "parallel",
       },
     },

docs/tools/plugin.md

@@ -147,10 +147,12 @@ such as `@beta` stay pinned to the selected package and fail when incompatible.
 
 Configure `security.installPolicy` to run a trusted local policy command before
 plugin install or update proceeds. The policy receives metadata plus the staged
-source path and can allow or block the install. It runs before plugin
-`before_install` hooks. The deprecated `--dangerously-force-unsafe-install`
-flag is accepted for compatibility but does not bypass install policy, hooks, or
-OpenClaw's built-in plugin dependency denylist.
+source path and can allow or block the install. It covers CLI and Gateway-backed
+plugin install/update paths. Plugin `before_install` hooks run later only in
+OpenClaw processes where plugin hooks are loaded, so use `security.installPolicy`
+for operator-owned install decisions. The deprecated
+`--dangerously-force-unsafe-install` flag is accepted for compatibility but does
+not bypass install policy or OpenClaw's built-in plugin dependency denylist.
 
 See [Skills config](/tools/skills-config#operator-install-policy-securityinstallpolicy)
 for the shared `security.installPolicy` exec schema used by both skills and

docs/tools/searxng-search.md

@@ -120,9 +120,9 @@ key wins first).
 - **Network guard** -- private/internal SearXNG endpoints opt in to
   private-network access; public `https://` SearXNG endpoints keep strict SSRF
   protection
-- **Auto-detection order** -- SearXNG is checked last (order 200) in
-  auto-detection. API-backed providers with configured keys run first, then
-  DuckDuckGo (order 100), then Ollama Web Search (order 110)
+- **Auto-detection order** -- SearXNG is checked after API-backed providers
+  with configured keys (order 200). Key-free providers such as DuckDuckGo or
+  Ollama Web Search are not auto-selected without an explicit provider choice
 - **Self-hosted** -- you control the instance, queries, and upstream search engines
 - **Categories** default to `general` when not configured
 - **Category fallback** -- if a non-`general` category request succeeds but
@@ -137,5 +137,5 @@ key wins first).
 ## Related
 
 - [Web Search overview](/tools/web) -- all providers and auto-detection
-- [DuckDuckGo Search](/tools/duckduckgo-search) -- another key-free fallback
+- [DuckDuckGo Search](/tools/duckduckgo-search) -- another key-free provider
 - [Brave Search](/tools/brave-search) -- structured results with free tier

docs/tools/tool-search.md

@@ -16,9 +16,9 @@ search or dynamic-tools surface. Codex-native code mode, tool search, deferred
 dynamic tools, and nested tool calls are stable Codex harness surfaces and do
 not depend on `tools.toolSearch`.
 
-When enabled for OpenClaw runs, 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:
+When enabled for OpenClaw runs, 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");
@@ -49,8 +49,8 @@ run:
 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 OpenClaw code bridge or the structured fallback tools to the
-   model.
+6. Expose the OpenClaw code bridge, the structured fallback tools, or the
+   compact directory surface 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.
@@ -59,18 +59,26 @@ normal policy, approval, hook, logging, and result handling still apply.
 
 ## Modes
 
-`tools.toolSearch` has two model-facing modes:
+`tools.toolSearch` has three 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.
+- `directory`: exposes `tool_search`, `tool_describe`, and `tool_call` plus a
+  bounded prompt directory of available tool names and descriptions for
+  providers that should see tool names without every full schema. OpenClaw can
+  also expose a small bounded set of likely or required tool schemas directly
+  for the current turn.
 
-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.
+All modes use the same policy-filtered catalog and normal OpenClaw execution
+path. If the current runtime cannot launch the isolated Node code-mode child
+process, the default `code` mode falls back to `tools` before catalog
+compaction. In `directory` mode, client-provided tools stay directly visible
+for the current run while OpenClaw tools, plugin tools, and MCP tools can be
+compacted behind the directory catalog. A direct call to an exact hidden
+directory name is hydrated from that same authorized catalog before execution.
 
-Both modes are experimental. Prefer direct tool exposure for small OpenClaw tool
+All modes are experimental. Prefer direct tool exposure for small OpenClaw tool
 catalogs, and prefer the Codex-native stable surfaces for Codex harness runs.
 
 There is no separate source-selection config. When Tool Search is enabled, the
@@ -90,7 +98,10 @@ Tool Search changes the shape:
   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
+- Tool Search directory mode: the model sees a bounded directory plus
+  search/describe/call controls and a small bounded set of likely or required
+  schemas
+- during the turn: the model can load remaining schemas as needed
 
 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
@@ -132,6 +143,20 @@ The structured fallback mode exposes the same operations as tools:
 - `tool_describe`
 - `tool_call`
 
+Directory mode exposes:
+
+- `tool_search`
+- `tool_describe`
+- `tool_call`
+
+It also keeps client-provided tools directly visible and may expose a small
+bounded set of likely or required catalog tool schemas directly for the current
+turn. If the bounded directory omits entries, use `tool_search` to find them. If
+the model requests an exact hidden directory tool name directly, OpenClaw
+hydrates it from the authorized catalog before normal execution.
+Directory-mode client tool names must not collide with OpenClaw, plugin, or MCP
+tool names because exact deferred dispatch uses those names.
+
 ## Runtime boundary
 
 The code bridge runs in a short-lived Node subprocess. The subprocess starts
@@ -186,6 +211,18 @@ Use the structured fallback tools instead for OpenClaw runs:
 }
 ```
 
+Use the compact directory surface instead for OpenClaw runs:
+
+```json5
+{
+  tools: {
+    toolSearch: {
+      mode: "directory",
+    },
+  },
+}
+```
+
 Tune code-mode timeout and search result limits:
 
 ```json5

docs/tools/web.md

@@ -6,7 +6,7 @@ read_when:
   - You want to enable or configure web_search
   - You want to enable or configure x_search
   - You need to choose a search provider
-  - You want to understand auto-detection and provider fallback
+  - You want to understand auto-detection and provider selection
 ---
 
 The `web_search` tool searches the web using your configured provider and
@@ -60,8 +60,11 @@ local while `web_search` and `x_search` can use xAI Responses under the hood.
   <Card title="Brave Search" icon="shield" href="/tools/brave-search">
     Structured results with snippets. Supports `llm-context` mode, country/language filters. Free tier available.
   </Card>
+  <Card title="Codex Hosted Search" icon="search" href="/plugins/codex-harness">
+    AI-synthesized grounded answers through your Codex app-server account.
+  </Card>
   <Card title="DuckDuckGo" icon="bird" href="/tools/duckduckgo-search">
-    Key-free fallback. No API key needed. Unofficial HTML-based integration.
+    Key-free provider. No API key needed. Unofficial HTML-based integration.
   </Card>
   <Card title="Exa" icon="brain" href="/tools/exa-search">
     Neural + keyword search with content extraction (highlights, text, summaries).
@@ -88,7 +91,7 @@ local while `web_search` and `x_search` can use xAI Responses under the hood.
     Paid Parallel Search API (`PARALLEL_API_KEY`); higher rate limits and objective tuning.
   </Card>
   <Card title="Parallel Search (Free)" icon="layer-group" href="/tools/parallel-search">
-    Zero-config default. Parallel's free Search MCP, with LLM-optimized dense excerpts and no API key.
+    Key-free opt-in. Parallel's free Search MCP, with LLM-optimized dense excerpts and no API key.
   </Card>
   <Card title="Perplexity" icon="search" href="/tools/perplexity-search">
     Structured results with content extraction controls and domain filtering.
@@ -106,6 +109,7 @@ local while `web_search` and `x_search` can use xAI Responses under the hood.
 | Provider                                         | Result style                                                   | Filters                                          | API key                                                                                 |
 | ------------------------------------------------ | -------------------------------------------------------------- | ------------------------------------------------ | --------------------------------------------------------------------------------------- |
 | [Brave](/tools/brave-search)                     | Structured snippets                                            | Country, language, time, `llm-context` mode      | `BRAVE_API_KEY`                                                                         |
+| [Codex Hosted Search](/plugins/codex-harness)    | AI-synthesized + source URLs                                   | Domains, context size, user location             | None; uses Codex/OpenAI sign-in                                                         |
 | [DuckDuckGo](/tools/duckduckgo-search)           | Structured snippets                                            | --                                               | None (key-free)                                                                         |
 | [Exa](/tools/exa-search)                         | Structured + extracted                                         | Neural/keyword mode, date, content extraction    | `EXA_API_KEY`                                                                           |
 | [Firecrawl](/tools/firecrawl)                    | Structured snippets                                            | Via `firecrawl_search` tool                      | `FIRECRAWL_API_KEY`                                                                     |
@@ -194,15 +198,16 @@ API-backed providers first:
 9. **Tavily** -- `TAVILY_API_KEY` or `plugins.entries.tavily.config.webSearch.apiKey` (order 70)
 10. **Parallel** -- paid Parallel Search API via `PARALLEL_API_KEY` or `plugins.entries.parallel.config.webSearch.apiKey`; optional `plugins.entries.parallel.config.webSearch.baseUrl` overrides the endpoint (order 75)
 
-Key-free fallbacks after that:
+Configured endpoint providers after that:
 
-11. **Parallel Search (Free)** -- the zero-config default: works with no account or API key via Parallel's free hosted [Search MCP](https://docs.parallel.ai/integrations/mcp/search-mcp) (order 76)
-12. **DuckDuckGo** -- key-free HTML fallback with no account or API key (order 100)
-13. **Ollama Web Search** -- key-free fallback via your configured local Ollama host when it is reachable and signed in with `ollama signin`; can reuse Ollama provider bearer auth when the host needs it, and can call direct `https://ollama.com` search when configured with `OLLAMA_API_KEY` (order 110)
-14. **SearXNG** -- `SEARXNG_BASE_URL` or `plugins.entries.searxng.config.webSearch.baseUrl` (order 200)
+11. **SearXNG** -- `SEARXNG_BASE_URL` or `plugins.entries.searxng.config.webSearch.baseUrl` (order 200)
 
-When no API-backed provider is configured, OpenClaw defaults to **Parallel
-Search (Free)**, so `web_search` works without an API key.
+Key-free providers such as **Parallel Search (Free)**, **DuckDuckGo**,
+**Ollama Web Search**, and **Codex Hosted Search** are available only when you
+select them explicitly with `tools.web.search.provider` or through
+`openclaw configure --section web`. OpenClaw does not send managed
+`web_search` queries to a key-free provider just because no API-backed provider
+is configured.
 
 OpenAI Responses models are an exception: while `tools.web.search.provider` is
 unset, they use OpenAI's native web search instead of the managed providers

extensions/acpx/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "@openclaw/acpx",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@openclaw/acpx",
-      "version": "2026.6.2",
+      "version": "2026.6.8",
       "dependencies": {
         "@agentclientprotocol/claude-agent-acp": "0.39.0",
         "@zed-industries/codex-acp": "0.15.0",
@@ -1548,9 +1548,9 @@
       }
     },
     "node_modules/hono": {
-      "version": "4.12.21",
-      "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.21.tgz",
-      "integrity": "sha512-uV63apnb0kyPtAUwoWgaGh9HyIFcv8lgmzPZSiTBQAFOFGIzka5EZ1dZocmGnn0XdX0+XTqJ6Tqv7selMuGLRQ==",
+      "version": "4.12.25",
+      "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.25.tgz",
+      "integrity": "sha512-2NFaIyNVgJmBs/ecmtGzlmluTFs5cHEWGTdu0t1HBwYzoGXOL5nUQBRMXsXWla5i4KkG//QMzVP88m1+I3fdAQ==",
       "license": "MIT",
       "engines": {
         "node": ">=16.9.0"

extensions/acpx/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/acpx",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "description": "OpenClaw ACP runtime backend with plugin-owned session and transport management.",
   "repository": {
     "type": "git",
@@ -26,10 +26,10 @@
       "minHostVersion": ">=2026.4.25"
     },
     "compat": {
-      "pluginApi": ">=2026.6.2"
+      "pluginApi": ">=2026.6.8"
     },
     "build": {
-      "openclawVersion": "2026.6.2",
+      "openclawVersion": "2026.6.8",
       "staticAssets": [
         {
           "source": "./src/runtime-internals/mcp-proxy.mjs",

extensions/active-memory/openclaw.plugin.json

@@ -123,11 +123,12 @@
       "help": "Optional explicit denylist of chat/user IDs. Sessions whose resolved conversation id matches the list are skipped even when the chat type is allowed. Applied after allowedChatIds."
     },
     "timeoutMs": {
-      "label": "Timeout (ms)"
+      "label": "Timeout (ms)",
+      "help": "Recall work budget on the main lane. Before recall, the hook allows up to 1500 ms for session/config preflight. After recall starts, it reserves another fixed 1500 ms only for abort settlement and transcript recovery."
     },
     "setupGraceTimeoutMs": {
       "label": "Setup Grace Timeout (ms)",
-      "help": "Advanced: extra blocking budget for cold embedded-run setup before the recall timeout is considered exhausted. Defaults to 0 so timeoutMs remains the main-lane hook budget unless you opt in."
+      "help": "Advanced: extra recall-work budget for cold embedded-run setup. Defaults to 0. The separate 1500 ms preflight cap and 1500 ms post-recall completion allowance still apply."
     },
     "queryMode": {
       "label": "Query Mode",

extensions/admin-http-rpc/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/admin-http-rpc",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "private": true,
   "description": "OpenClaw admin HTTP RPC endpoint",
   "type": "module",

extensions/alibaba/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/alibaba-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "private": true,
   "description": "OpenClaw Alibaba Model Studio video provider plugin",
   "type": "module",

extensions/amazon-bedrock-mantle/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "@openclaw/amazon-bedrock-mantle-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@openclaw/amazon-bedrock-mantle-provider",
-      "version": "2026.6.2",
+      "version": "2026.6.8",
       "dependencies": {
         "@anthropic-ai/sdk": "0.100.1",
         "@aws/bedrock-token-generator": "1.1.0"

extensions/amazon-bedrock-mantle/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/amazon-bedrock-mantle-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "description": "OpenClaw Amazon Bedrock Mantle provider plugin for OpenAI-compatible model routing.",
   "repository": {
     "type": "git",
@@ -24,10 +24,10 @@
       "minHostVersion": ">=2026.5.12-beta.1"
     },
     "compat": {
-      "pluginApi": ">=2026.6.2"
+      "pluginApi": ">=2026.6.8"
     },
     "build": {
-      "openclawVersion": "2026.6.2",
+      "openclawVersion": "2026.6.8",
       "bundledDist": false
     },
     "release": {

extensions/amazon-bedrock/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "@openclaw/amazon-bedrock-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@openclaw/amazon-bedrock-provider",
-      "version": "2026.6.2",
+      "version": "2026.6.8",
       "dependencies": {
         "@aws-sdk/client-bedrock": "3.1056.0",
         "@aws-sdk/client-bedrock-runtime": "3.1056.0",

extensions/amazon-bedrock/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/amazon-bedrock-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "description": "OpenClaw Amazon Bedrock provider plugin with model discovery, embeddings, and guardrail support.",
   "repository": {
     "type": "git",
@@ -28,10 +28,10 @@
       "minHostVersion": ">=2026.5.12-beta.1"
     },
     "compat": {
-      "pluginApi": ">=2026.6.2"
+      "pluginApi": ">=2026.6.8"
     },
     "build": {
-      "openclawVersion": "2026.6.2",
+      "openclawVersion": "2026.6.8",
       "bundledDist": false
     },
     "release": {

extensions/anthropic-vertex/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "@openclaw/anthropic-vertex-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@openclaw/anthropic-vertex-provider",
-      "version": "2026.6.2",
+      "version": "2026.6.8",
       "dependencies": {
         "@anthropic-ai/vertex-sdk": "0.16.1"
       }

extensions/anthropic-vertex/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/anthropic-vertex-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "description": "OpenClaw Anthropic Vertex provider plugin for Claude models on Google Vertex AI.",
   "repository": {
     "type": "git",
@@ -23,10 +23,10 @@
       "minHostVersion": ">=2026.5.12-beta.1"
     },
     "compat": {
-      "pluginApi": ">=2026.6.2"
+      "pluginApi": ">=2026.6.8"
     },
     "build": {
-      "openclawVersion": "2026.6.2",
+      "openclawVersion": "2026.6.8",
       "bundledDist": false
     },
     "release": {

extensions/anthropic/cli-backend.ts

@@ -34,6 +34,7 @@ export function buildAnthropicCliBackend(): CliBackendPlugin {
     bundleMcp: true,
     bundleMcpMode: "claude-config-file",
     nativeToolMode: "always-on",
+    sideQuestionToolMode: "disabled",
     ownsNativeCompaction: true,
     config: {
       command: "claude",

extensions/anthropic/cli-shared.test.ts

@@ -150,6 +150,61 @@ describe("resolveClaudeCliExecutionArgs", () => {
       }),
     ).toEqual(["-p", "--effort", "max"]);
   });
+
+  it("forces isolated no-tool one-shot args for side-question execution", () => {
+    expect(
+      resolveClaudeCliExecutionArgs({
+        workspaceDir: "/tmp",
+        provider: "claude-cli",
+        modelId: "claude-opus-4-7",
+        thinkingLevel: "max",
+        useResume: true,
+        executionMode: "side-question",
+        baseArgs: [
+          "-p",
+          "--output-format",
+          "stream-json",
+          "--allowedTools=mcp__openclaw__*",
+          "--allowedTools",
+          "Read",
+          "Grep",
+          "--permission-mode",
+          "bypassPermissions",
+          "--session-id=abc",
+          "--resume",
+          "old-session",
+          "--resume-session-at",
+          "old-message",
+          "--resume-session-at=old-message-equals",
+          "--mcp-config",
+          "/tmp/side-question-mcp.json",
+          "--bare",
+          "--safe-mode",
+          "--strict-mcp-config",
+          "--no-session-persistence",
+          "--max-turns",
+          "4",
+          "--effort",
+          "high",
+        ],
+      }),
+    ).toEqual([
+      "-p",
+      "--output-format",
+      "stream-json",
+      "--safe-mode",
+      "--tools",
+      "",
+      "--disallowedTools",
+      "mcp__*",
+      "--strict-mcp-config",
+      "--no-session-persistence",
+      "--max-turns",
+      "1",
+      "--permission-mode",
+      "default",
+    ]);
+  });
 });
 
 describe("normalizeClaudeBackendConfig", () => {

extensions/anthropic/cli-shared.ts

@@ -67,8 +67,26 @@ const CLAUDE_LEGACY_SKIP_PERMISSIONS_ARG = "--dangerously-skip-permissions";
 const CLAUDE_PERMISSION_MODE_ARG = "--permission-mode";
 const CLAUDE_SETTING_SOURCES_ARG = "--setting-sources";
 const CLAUDE_EFFORT_ARG = "--effort";
+const CLAUDE_BARE_ARG = "--bare";
+const CLAUDE_SAFE_MODE_ARG = "--safe-mode";
+const CLAUDE_TOOLS_ARG = "--tools";
+const CLAUDE_DISALLOWED_TOOLS_ARG = "--disallowedTools";
+const CLAUDE_MCP_CONFIG_ARG = "--mcp-config";
+const CLAUDE_STRICT_MCP_CONFIG_ARG = "--strict-mcp-config";
+const CLAUDE_NO_SESSION_PERSISTENCE_ARG = "--no-session-persistence";
+const CLAUDE_MAX_TURNS_ARG = "--max-turns";
+const CLAUDE_SESSION_ID_ARG = "--session-id";
+const CLAUDE_RESUME_ARG = "--resume";
+const CLAUDE_RESUME_SESSION_AT_ARG = "--resume-session-at";
+const CLAUDE_RESUME_SHORT_ARG = "-r";
+const CLAUDE_CONTINUE_ARG = "--continue";
+const CLAUDE_CONTINUE_SHORT_ARG = "-c";
+const CLAUDE_FORK_SESSION_ARG = "--fork-session";
 const CLAUDE_SAFE_SETTING_SOURCES = "user";
 const CLAUDE_BYPASS_PERMISSION_MODE = "bypassPermissions";
+const CLAUDE_DEFAULT_PERMISSION_MODE = "default";
+const CLAUDE_NO_TOOLS_VALUE = "";
+const CLAUDE_DENY_MCP_TOOLS_VALUE = "mcp__*";
 
 type ClaudeCliEffort = "low" | "medium" | "high" | "xhigh" | "max";
 
@@ -232,10 +250,89 @@ function stripClaudeEffortArgs(args: readonly string[]): string[] {
   return normalized;
 }
 
+const CLAUDE_SIDE_QUESTION_VARIADIC_VALUE_ARGS = new Set([
+  "--allowedTools",
+  "--allowed-tools",
+  CLAUDE_DISALLOWED_TOOLS_ARG,
+  "--disallowed-tools",
+  CLAUDE_TOOLS_ARG,
+  CLAUDE_MCP_CONFIG_ARG,
+]);
+
+const CLAUDE_SIDE_QUESTION_VALUE_ARGS = new Set([
+  CLAUDE_PERMISSION_MODE_ARG,
+  CLAUDE_SESSION_ID_ARG,
+  CLAUDE_RESUME_ARG,
+  CLAUDE_RESUME_SESSION_AT_ARG,
+  CLAUDE_RESUME_SHORT_ARG,
+  CLAUDE_MAX_TURNS_ARG,
+]);
+
+const CLAUDE_SIDE_QUESTION_BARE_ARGS = new Set([
+  CLAUDE_CONTINUE_ARG,
+  CLAUDE_CONTINUE_SHORT_ARG,
+  CLAUDE_FORK_SESSION_ARG,
+  CLAUDE_BARE_ARG,
+  CLAUDE_SAFE_MODE_ARG,
+  CLAUDE_STRICT_MCP_CONFIG_ARG,
+  CLAUDE_NO_SESSION_PERSISTENCE_ARG,
+]);
+
+function stripClaudeSideQuestionConflictingArgs(args: readonly string[]): string[] {
+  const normalized: string[] = [];
+  for (let i = 0; i < args.length; i += 1) {
+    const arg = args[i] ?? "";
+    const equalsIndex = arg.indexOf("=");
+    const argName = equalsIndex > 0 ? arg.slice(0, equalsIndex) : arg;
+    if (CLAUDE_SIDE_QUESTION_BARE_ARGS.has(argName)) {
+      continue;
+    }
+    if (CLAUDE_SIDE_QUESTION_VARIADIC_VALUE_ARGS.has(argName)) {
+      if (equalsIndex < 0) {
+        while (typeof args[i + 1] === "string" && !args[i + 1]?.startsWith("-")) {
+          i += 1;
+        }
+      }
+      continue;
+    }
+    if (CLAUDE_SIDE_QUESTION_VALUE_ARGS.has(argName)) {
+      if (equalsIndex < 0) {
+        const maybeValue = args[i + 1];
+        if (typeof maybeValue === "string" && !maybeValue.startsWith("-")) {
+          i += 1;
+        }
+      }
+      continue;
+    }
+    normalized.push(arg);
+  }
+  return normalized;
+}
+
+function resolveClaudeCliSideQuestionExecutionArgs(baseArgs: readonly string[]): string[] {
+  return [
+    ...stripClaudeSideQuestionConflictingArgs(stripClaudeEffortArgs(baseArgs)),
+    CLAUDE_SAFE_MODE_ARG,
+    CLAUDE_TOOLS_ARG,
+    CLAUDE_NO_TOOLS_VALUE,
+    CLAUDE_DISALLOWED_TOOLS_ARG,
+    CLAUDE_DENY_MCP_TOOLS_VALUE,
+    CLAUDE_STRICT_MCP_CONFIG_ARG,
+    CLAUDE_NO_SESSION_PERSISTENCE_ARG,
+    CLAUDE_MAX_TURNS_ARG,
+    "1",
+    CLAUDE_PERMISSION_MODE_ARG,
+    CLAUDE_DEFAULT_PERMISSION_MODE,
+  ];
+}
+
 /** Resolve final Claude CLI execution args for one backend invocation. */
 export function resolveClaudeCliExecutionArgs(
   context: CliBackendResolveExecutionArgsContext,
 ): string[] {
+  if (context.executionMode === "side-question") {
+    return resolveClaudeCliSideQuestionExecutionArgs(context.baseArgs);
+  }
   const effort = mapClaudeCliThinkingLevelToEffort(context.thinkingLevel);
   if (!effort) {
     return [...context.baseArgs];

extensions/anthropic/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/anthropic-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "private": true,
   "description": "OpenClaw Anthropic provider plugin",
   "type": "module",

extensions/arcee/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/arcee-provider",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "private": true,
   "description": "OpenClaw Arcee provider plugin",
   "type": "module",

extensions/azure-speech/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/azure-speech",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "private": true,
   "description": "OpenClaw Azure Speech plugin",
   "type": "module",

extensions/bonjour/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/bonjour",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "description": "OpenClaw Bonjour/mDNS gateway discovery",
   "type": "module",
   "dependencies": {

extensions/brave/npm-shrinkwrap.json

@@ -1,12 +1,12 @@
 {
   "name": "@openclaw/brave-plugin",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@openclaw/brave-plugin",
-      "version": "2026.6.2"
+      "version": "2026.6.8"
     }
   }
 }

extensions/brave/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/brave-plugin",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "description": "OpenClaw Brave Search provider plugin for web search.",
   "repository": {
     "type": "git",
@@ -21,10 +21,10 @@
       "allowInvalidConfigRecovery": true
     },
     "compat": {
-      "pluginApi": ">=2026.6.2"
+      "pluginApi": ">=2026.6.8"
     },
     "build": {
-      "openclawVersion": "2026.6.2"
+      "openclawVersion": "2026.6.8"
     },
     "release": {
       "publishToClawHub": true,

extensions/browser/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/browser-plugin",
-  "version": "2026.6.2",
+  "version": "2026.6.8",
   "private": true,
   "description": "OpenClaw browser tool plugin",
   "type": "module",

extensions/browser/skills/browser-automation/SKILL.md

@@ -25,7 +25,7 @@ Use this skill when you need the `browser` tool for anything beyond a single pag
    - Use the same `targetId` for follow-up actions so refs stay on the same tab.
    - For durable Playwright refs, request `refs="aria"` when supported. If you receive `axN` refs from `snapshotFormat="aria"`, use them only after that same snapshot call; stale or unbound `axN` refs fail fast and need a fresh snapshot.
    - Use `urls=true` when link text is ambiguous or a direct navigation target would avoid brittle clicks.
-   - Use `labels=true` on snapshot or screenshot when visual position matters.
+   - Use `labels=true` on snapshot or screenshot when visual position matters. On Playwright-backed profiles, the response includes an `annotations` array (`{ref, number, role, name?, box}`) with each ref's bounding box in the captured image's coordinate space, so you can reason about position without re-snapshotting; screenshot labels can also combine with `fullPage=true` (CLI: `--full-page`) to label the whole document, or `ref` / `element` to clip to one element. `profile="user"` and other existing-session (chrome-mcp) profiles render an overlay into page screenshots but do not attach `annotations` or use the Playwright full-page/ref/element projection helper, so read positions from the labeled image itself on those profiles. The raw-CDP fallback (no Playwright) does not support labeled screenshots at all and returns a 501, so only request `labels` when Playwright is available.
 4. Act narrowly:
    - Prefer `action="act"` with a ref from the latest snapshot.
    - After navigation, modal changes, or form submission, snapshot again before the next action.

extensions/browser/src/browser-tool.actions.ts

@@ -486,6 +486,7 @@ export async function executeSnapshotAction(params: {
       labels: snapshot.labels,
       labelsCount: snapshot.labelsCount,
       labelsSkipped: snapshot.labelsSkipped,
+      annotations: snapshot.annotations,
       imagePath: snapshot.imagePath,
       imageType: snapshot.imageType,
       refsFallback,

extensions/browser/src/browser/client-actions-types.ts

@@ -1,6 +1,8 @@
 /**
  * Shared result types for browser client action helpers.
  */
+import type { AnnotationItem } from "./screenshot-annotate.js";
+
 /** Generic success result for action endpoints. */
 export type BrowserActionOk = { ok: true };
 
@@ -20,4 +22,10 @@ export type BrowserActionPathResult = {
   labels?: boolean;
   labelsCount?: number;
   labelsSkipped?: number;
+  /**
+   * Per-ref bounding boxes when labels=true. Coordinates are in the
+   * captured image's space (viewport / fullpage / element-relative).
+   * Omitted when empty.
+   */
+  annotations?: AnnotationItem[];
 };

extensions/browser/src/browser/client.ts

@@ -18,6 +18,7 @@ import type {
 } from "./client.types.js";
 import { DEFAULT_BROWSER_SNAPSHOT_TIMEOUT_MS } from "./constants.js";
 import type { BrowserDoctorReport } from "./doctor.js";
+import type { AnnotationItem } from "./screenshot-annotate.js";
 
 export type { BrowserStatus, BrowserTab, BrowserTransport } from "./client.types.js";
 export type { BrowserDoctorCheck, BrowserDoctorReport } from "./doctor.js";
@@ -124,6 +125,11 @@ export type SnapshotResult =
       labels?: boolean;
       labelsCount?: number;
       labelsSkipped?: number;
+      /**
+       * Per-ref bounding boxes when labels=true. Coordinates are in the
+       * captured image's space. Omitted when empty.
+       */
+      annotations?: AnnotationItem[];
       imagePath?: string;
       imageType?: "png" | "jpeg";
       blockedByDialog?: boolean;