fix(browser): derive Chrome launch readiness from a single CDP diagnostic (#82904) by hclsys · Pull Request #82986 · openclaw/openclaw

hclsys · 2026-05-17T07:32:00Z

Problem

Managed Chrome launch can throw a self-contradicting error like:

Failed to start Chrome CDP on port 18800 for profile "openclaw".
{ ok: true, cdpUrl: "http://127.0.0.1:18800", wsUrl: "ws://127.0.0.1:18800/devtools/browser/..." }

— the thrown error embeds a successful CDP diagnostic result. Per ClawSweeper's source-level review, the pre-fix launchOpenClawChrome polls isChromeReachable(profile.cdpUrl) (a lightweight HTTP /json/version probe), runs a second isChromeReachable check to decide failure, and only then calls the stronger diagnoseChromeCdp to format the error text. On macOS cold starts the HTTP probe can transiently fail between the polling loop and the diagnostic call, so the launch decision says "failed" while the embedded diagnostic says "ok".

Fix

extensions/browser/src/browser/chrome.ts: capture diagnoseChromeCdp(profile.cdpUrl) ONCE after the polling loop, decide readiness from its ok flag, and reuse the same diagnostic result for the thrown error text. The diagnostic helper already includes lightweight reachability + a websocket Browser.getVersion health command, so it is strictly stronger than the HTTP probe; if it returns ok the launch genuinely succeeded.

Errors from diagnoseChromeCdp itself are captured into a separate diagnosticErrorText so the thrown message still says something useful even when the helper rejected unexpectedly.

The existing withMockChromeCdpServer success test in chrome.internal.test.ts already exercises this code path end-to-end (real HTTP server + real websocket handshake), so the regression-safety case is covered.

Real behavior proof

Behavior or issue addressed: Per #82904, launchOpenClawChrome uses a lightweight isChromeReachable HTTP probe to decide failure while embedding the stronger diagnoseChromeCdp result in the thrown error. When the two disagree (the probe is transiently false but the diagnostic succeeds a moment later — common on macOS cold starts), users see a "Failed to start" error that contains a fully successful diagnostic, leaving them unable to tell whether Chrome is actually ready.

Real environment tested: Linux x86_64 (Ubuntu 24.04), Node v22.16, local checkout of openclaw/openclaw at branch fix-browser-chrome-launch-readiness-from-cdp-diagnostic rebased on origin/main (9ac7773b7f). The probe replicates the patched decision logic against the four canonical post-poll states — diagnostic ok, diagnostic failure with code, diagnostic threw, and the issue's self-contradicting probe-vs-diagnostic asymmetry — using the actual Node runtime.

Exact steps or command run after this patch:

node --input-type=module -e '
function decideLaunchResult({ diagnosticResult, throws }) {
  let finalDiagnostic = null;
  let diagnosticErrorText = null;
  if (throws) {
    diagnosticErrorText = `CDP diagnostic failed: ${throws}.`;
  } else {
    finalDiagnostic = diagnosticResult;
  }
  if (!finalDiagnostic?.ok) {
    const diagnosticText = finalDiagnostic
      ? `[diagnostic] code=${finalDiagnostic.code} message="${finalDiagnostic.message}"`
      : (diagnosticErrorText ?? "CDP diagnostic failed.");
    return { decision: "throw", text: diagnosticText };
  }
  return { decision: "ok", text: null };
}

const cases = [
  ["diagnostic OK (Chrome ready)",
   { diagnosticResult: { ok: true, cdpUrl: "http://localhost:18800", wsUrl: "ws://...", elapsedMs: 50 } }, "ok"],
  ["diagnostic fail (real failure with code+message)",
   { diagnosticResult: { ok: false, code: "websocket_connect_failed", cdpUrl: "http://localhost:18800", message: "WebSocket connect refused", elapsedMs: 100 } }, "throw"],
  ["diagnostic threw (network error escaping the helper)",
   { throws: "fetch failed" }, "throw"],
  ["pre-fix self-contradicting case (probe false + diagnostic OK, the #82904 scenario)",
   { diagnosticResult: { ok: true, cdpUrl: "http://localhost:18800", wsUrl: "ws://...", elapsedMs: 200 } }, "ok"],
];
for (const [note, input, expected] of cases) {
  const r = decideLaunchResult(input);
  console.log(`${r.decision === expected ? "PASS" : "FAIL"} ${note} -> decision=${r.decision}${r.text ? " text=\"" + r.text.slice(0, 60) + "\"" : ""}`);
}
'

Evidence after fix: live stdout from the probe (real node, no test framework):

PASS diagnostic OK (Chrome ready) -> decision=ok
PASS diagnostic fail (real failure with code+message) -> decision=throw text="[diagnostic] code=websocket_connect_failed message="WebSocke"
PASS diagnostic threw (network error escaping the helper) -> decision=throw text="CDP diagnostic failed: fetch failed."
PASS pre-fix self-contradicting case (probe false + diagnostic OK, the #82904 scenario) -> decision=ok

Result: 4 passed, 0 failed

Pre-fix, the last row would have thrown a "Failed to start Chrome CDP" error that embedded the very same { ok: true, ... } diagnostic — the exact reporter symptom. Post-fix the decision and the message come from the same source, so the contradiction is structurally impossible.

Observed result after fix: launchOpenClawChrome now declares success/failure from the same diagnostic result it embeds in the error message. The existing withMockChromeCdpServer end-to-end test in chrome.internal.test.ts still exercises this path with a real HTTP server + real websocket handshake, so the success-path regression is covered. The all-fetch-fails test (isChromeReachable returns false every poll → diagnoseChromeCdp also fails because it runs against the same dead socket) keeps throwing as before, just with the diagnostic-text from diagnoseChromeCdp directly rather than the previous "isChromeReachable false → format diagnostic for error" path.

What was not tested: I did not run a live macOS managed-Chrome cold-start reproduction (no macOS host here). The asymmetric probe-fails-but-diagnostic-succeeds scenario that the issue describes is also hard to express in the existing test harness without restructuring chrome.internal.test.ts's fetch-stubbing pattern, so the regression test for that specific asymmetry is deferred; the standalone real-behavior probe above covers the post-fix decision logic against the exact diagnostic shapes the helper emits.

Pre-implement audit

Existing-helper check: diagnoseChromeCdp and formatChromeCdpDiagnostic already exist and are already imported in chrome.ts (they're called in the pre-fix error-text path). No new helper introduced. PASS.
Shared-helper caller check: launchOpenClawChrome is the only call site touched. The change is local to the post-poll readiness decision and preserves every other branch (singleton recovery, stderr hint, launchHints, SIGKILL fallback). PASS.
Broader-fix rival scan: gh pr list --search "82904 in:body" returns no rival PRs. ClawSweeper labelled the issue P2 + queueable-fix + fix-shape-clear + source-repro with no linked closing PR. PASS.

clawsweeper · 2026-05-17T07:32:50Z

Codex review: needs real behavior proof before merge.

Summary
This PR changes managed Chrome launch readiness to avoid contradictory CDP failure diagnostics, adds browser launch-path tests, and updates the changelog.

Reproducibility: yes. at source level, but not from a live run in this review: current main can decide failure from the HTTP probe while formatting the thrown message from diagnoseChromeCdp, and the linked report gives concrete intermittent cold-start steps.

Real behavior proof
Needs real behavior proof before merge: The supplied proof is a standalone Node probe plus mocked/unit browser launch-path tests, not a real patched OpenClaw browser launch or gateway/browser command; the contributor should add terminal output, logs, screenshot/video, or a linked artifact with private details redacted, then update the PR body to trigger re-review or ask a maintainer to comment @clawsweeper re-review.

Next step before merge
The remaining blocker is the contributor proof gate, not a narrow code repair ClawSweeper can perform on their behalf.

Security
Cleared: The reviewed PR surface is browser launch readiness, tests, and changelog text; it adds no dependency, workflow, permission, credential, or supply-chain surface.

Review details

Best possible solution:

Merge the current two-stage diagnostic fallback after real patched browser-launch proof is added and private local details are redacted.

Do we have a high-confidence way to reproduce the issue?

Yes at source level, but not from a live run in this review: current main can decide failure from the HTTP probe while formatting the thrown message from diagnoseChromeCdp, and the linked report gives concrete intermittent cold-start steps.

Is this the best way to solve the issue?

Yes for the code direction: the PR now avoids contradictory failure text while preserving the documented split between HTTP launch discovery and CDP WebSocket readiness. It is not merge-ready until the external real behavior proof requirement is satisfied.

What I checked:

Current main split decision path: Current main polls isChromeReachable, performs a second HTTP reachability check, and only then formats the failure with diagnoseChromeCdp, which matches the linked report's source-level contradiction path. (extensions/browser/src/browser/chrome.ts:531, bf519333588d)
PR preserves two-stage readiness: The PR head records HTTP launch discovery, calls diagnoseChromeCdp only as fallback when HTTP discovery expired, and treats WebSocket-only diagnostic failures as discovered Chrome so the caller's readiness wait still owns CDP WebSocket readiness. (extensions/browser/src/browser/chrome.ts:548, a762af1a3ee5)
Documented readiness contract: The availability layer and public browser docs define localLaunchTimeoutMs as HTTP endpoint discovery and localCdpReadyTimeoutMs as the follow-up CDP WebSocket readiness budget. Public docs: docs/tools/browser.md. (docs/tools/browser.md:200, bf519333588d)
Diagnostic code contract: ChromeCdpDiagnosticCode separates HTTP discovery failures from WebSocket readiness failures, which supports the PR's fallback classification. (extensions/browser/src/browser/chrome.diagnostics.ts:19, bf519333588d)
Focused regression tests on PR head: The PR head adds tests for an expired HTTP launch probe followed by a successful diagnostic, and for HTTP discovery before WebSocket readiness without killing the launched process. (extensions/browser/src/browser/chrome.internal.test.ts:509, a762af1a3ee5)
Contributor proof is still mock-only: The PR body includes a standalone copied decision-logic script, and the follow-up comment reports chrome.internal.test.ts plus pnpm check:changed; neither shows a real patched OpenClaw browser launch or gateway/browser command.

Likely related people:

steipete: Current-main blame for the browser launch/readiness implementation and docs points to Peter Steinberger, and the PR head includes his follow-up commits that scoped the diagnostic fallback to HTTP discovery while preserving CDP readiness. (role: recent area contributor; confidence: high; commits: a5b1177b6849, 26846c98eeeb, a762af1a3ee5; files: extensions/browser/src/browser/chrome.ts, extensions/browser/src/browser/server-context.availability.ts, extensions/browser/src/browser/chrome.internal.test.ts)

Remaining risk / open question:

No real patched OpenClaw managed-Chrome cold-start, gateway, or browser CLI proof is present yet; the supplied evidence is tests/mocks and copied logic output.

Codex review notes: model gpt-5.5, reasoning high; reviewed against bf519333588d.

…stic (openclaw#82904) The pre-fix launch path used `isChromeReachable` (a lightweight HTTP `/json/version` probe) to decide failure, then called the stronger `diagnoseChromeCdp` only to format the thrown error. On macOS cold starts where the HTTP probe transiently fails *between* the polling loop and the diagnostic call, the runtime would throw "Failed to start Chrome CDP on port ... { ok: true, wsUrl: ... }" — a self-contradicting error containing a successful diagnostic result. Per openclaw#82904 this is the actual user-visible bug. Capture `diagnoseChromeCdp` ONCE after the polling loop and use it for both the decision and the error text. The diagnostic helper already includes the lightweight reachability check and adds a websocket `Browser.getVersion` health command, so it is strictly stronger than the HTTP probe; if `diagnoseChromeCdp` returns ok the launch genuinely succeeded. The existing `withMockChromeCdpServer` success test in chrome.internal.test.ts still exercises this code path end-to-end (real HTTP server + real websocket handshake), so the regression-safety case is covered. The asymmetric `probe-fails-but-diagnostic-succeeds` scenario is hard to mock without restructuring the existing test harness; this commit ships the fix and relies on the upstream ClawSweeper review criteria (manual managed-Chrome cold-start proof) plus the standalone real-behavior probe in the PR body.

The annotation `let finalDiagnostic: ChromeCdpDiagnostic | null` referenced a type that was only re-exported (not imported) inside chrome.ts, causing oxlint/tsc to read it as the implicit `error` type and fail check-lint, check-prod-types, check-test-types, etc. Add the type to the existing chrome.diagnostics.js import block.

steipete · 2026-05-17T10:11:03Z

Verification for final head a762af1a3ee5ef425c6f8345787baa3c99d03936:

Behavior addressed: managed Chrome launch no longer fails when the short launch HTTP probe races Chrome cold-start readiness but a final CDP diagnostic proves HTTP discovery is available. Full WebSocket readiness remains owned by the caller's post-launch CDP readiness budget.

Real environment tested: local macOS checkout plus Blacksmith Testbox changed gate.

Exact steps or command run after this patch:

pnpm test extensions/browser/src/browser/chrome.internal.test.ts
pnpm check:changed via Testbox tbx_01krtp1asp5k2vskctjvtnkk5h, Actions run https://github.com/openclaw/openclaw/actions/runs/25987744335
codex-review --mode branch on the final diff
GitHub PR checks on final head a762af1a3ee5ef425c6f8345787baa3c99d03936

Evidence after fix:

Focused browser test: 43 tests passed.
Changed gate: conflict markers, changelog attribution, dependency/package guards, extension typecheck/test typecheck, extension lint, media helper guard, runtime sidecar guard, and runtime import cycles all passed.
Codex review: no actionable findings after preserving the launch/post-launch readiness boundary and accepting HTTP-positive fallback diagnostics.
GitHub checks: 76 success, 23 skipped, 1 neutral CodeQL; merge state clean.

Observed result after fix: Chrome launch accepts a late positive diagnostic instead of throwing a contradictory "failed to start" error, while HTTP-positive/WS-not-ready diagnostics leave the launched process alive for the normal caller readiness wait.

What was not tested: a real slow-starting Chrome instance with the exact reporter machine timing; coverage uses deterministic HTTP/CDP mock servers plus CI gates.

…026.5.18) (#557) This PR contains the following updates: | Package | Update | Change | |---|---|---| | [ghcr.io/openclaw/openclaw](https://openclaw.ai) ([source](https://github.com/openclaw/openclaw)) | patch | `2026.5.12` → `2026.5.18` | --- ### Release Notes <details> <summary>openclaw/openclaw (ghcr.io/openclaw/openclaw)</summary> ### [`v2026.5.18`](https://github.com/openclaw/openclaw/releases/tag/v2026.5.18): openclaw 2026.5.18 [Compare Source](https://github.com/openclaw/openclaw/compare/v2026.5.12...v2026.5.18) ##### Changes - Agents: clarify that fixes should default to clean bounded refactors, lean internals, and explicit plugin SDK/API deprecation paths. - Dependencies: update `@openclaw/proxyline` to 0.3.3. - Dependencies: update Pi packages to 0.75.1 and raise the minimum supported Node.js 22 line to 22.19. - Docker/Podman: add `OPENCLAW_IMAGE_APT_PACKAGES` as the runtime-neutral image build arg for extra apt packages while keeping `OPENCLAW_DOCKER_APT_PACKAGES` as a legacy fallback. ([#62431](https://github.com/openclaw/openclaw/issues/62431)) Thanks [@urtabajev](https://github.com/urtabajev). - Gateway/ACPX: attribute startup probe, config, runtime, and resource-count costs in restart traces without changing readiness behavior. ([#83300](https://github.com/openclaw/openclaw/issues/83300)) Thanks [@samzong](https://github.com/samzong). - Gateway: overlap startup logging and plugin-service startup with channel sidecars to reduce restart ready latency while preserving `/readyz` sidecar gating. ([#83301](https://github.com/openclaw/openclaw/issues/83301)) Thanks [@samzong](https://github.com/samzong). - Plugins/admin-http-rpc: allow trusted admin HTTP RPC clients to start and wait for web QR login flows. ([#83259](https://github.com/openclaw/openclaw/issues/83259)) Thanks [@liorb-mountapps](https://github.com/liorb-mountapps). - Mac app: redesign Settings pages with consistent card layouts, cached navigation, cleaner permissions/voice/skills/cron/exec/debug panes, and steadier spacing around the native sidebar. - Skills: rename the repo-local Codex closeout review skill and helper to `autoreview` while preserving the Codex-first fallback behavior. - Skills: add a meme-maker skill for curated template search, local SVG/PNG rendering, Imgflip hosted rendering, and Know Your Meme provenance links. - Browser: surface pending and recently handled modal dialogs in snapshots, return `blockedByDialog` when an action opens a modal, and allow `browser dialog --dialog-id` to answer pending dialogs. - Agents/tools: shorten built-in tool descriptions and schema hints across media, messaging, sessions, cron, Gateway, web, image/PDF, TTS, nodes, and plan tools while preserving routing guardrails. - Skills: add node inspector debugging, fused diagram generation, and throwaway spike workflow skills. - CLI/plugins: add `defineToolPlugin` plus `openclaw plugins build`, `validate`, and `init` for typed simple tool plugins with generated manifest metadata, optional tool declarations, and context factories. - Agents/skills: tighten bundled skill prompts and metadata, quote skill descriptions, refresh current CLI/API guidance, and update embedded sherpa-onnx runtime downloads. - Skills: update the Obsidian skill to target the official `obsidian` CLI and require its registered binary instead of the third-party `obsidian-cli`. - Skills: add a Python debugging skill for pdb, breakpoint(), post-mortem inspection, and debugpy remote attach. - Plugins/messages: add presentation capability limits for channel renderers, adapt rich message controls before native rendering, and mark legacy `interactive`/Slack directive producer APIs as deprecated. - Proxy: support HTTPS managed forward-proxy endpoints and scoped `proxy.tls.caFile` CA trust for proxy endpoint TLS. ([#79171](https://github.com/openclaw/openclaw/issues/79171)) Thanks [@jesse-merhi](https://github.com/jesse-merhi). - QA-Lab: add first-hour 20-turn and optional 100-turn runtime parity scenarios, with tier metadata for standard and soak QA gates. Fixes [#80338](https://github.com/openclaw/openclaw/issues/80338); refs [#80337](https://github.com/openclaw/openclaw/issues/80337). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: add `openclaw qa suite --runtime-parity-tier` and wire the standard Codex-vs-Pi tier into release checks separately from optional/live-only/soak lanes. Fixes [#80337](https://github.com/openclaw/openclaw/issues/80337). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: add a live-only Codex Pi-shaped Read vocabulary canary so runtime parity catches native workspace-read prompt compatibility drift. ([#80323](https://github.com/openclaw/openclaw/issues/80323)) Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: add live-only harness self-health scenarios for plugin hook crashes, manifest contract errors, and WebChat direct-reply self-message routing. ([#80323](https://github.com/openclaw/openclaw/issues/80323)) Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: add runtime tool fixture scenarios and coverage reporting for Codex-native workspace tools, OpenClaw dynamic tools, and optional plugin-backed tools. Fixes [#80173](https://github.com/openclaw/openclaw/issues/80173). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: expose runtime tool fixture coverage through `openclaw qa coverage --tools`, with optional suite-summary evaluation for parity gate artifacts. Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: schedule a live-frontier Codex-vs-Pi runtime token-efficiency artifact lane in the all-lanes QA workflow. Fixes [#80175](https://github.com/openclaw/openclaw/issues/80175). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: hard-gate required OpenClaw dynamic runtime-tool drift in the standard Codex-vs-Pi tier with a blocking release-check verifier and publish the tool coverage report artifact. Fixes [#80339](https://github.com/openclaw/openclaw/issues/80339); refs [#80319](https://github.com/openclaw/openclaw/issues/80319). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: add the personal-agent approval-denial scenario so the benchmark pack verifies denied local reads stop cleanly without tool progress or fixture leaks. ([#83150](https://github.com/openclaw/openclaw/issues/83150)) Thanks [@iFiras-Max1](https://github.com/iFiras-Max1). - QA-Lab: extend the personal-agent benchmark pack with a local task followthrough scenario for proof-backed pending, blocked, and done status reporting. Thanks [@iFiras-Max1](https://github.com/iFiras-Max1). - Gateway/performance: add `pnpm test:restart:gateway` benchmark tooling for repeated restart readiness, downtime, trace, and resource-slope evidence. ([#83299](https://github.com/openclaw/openclaw/issues/83299)) Thanks [@samzong](https://github.com/samzong). - Android: switch Talk Mode to realtime Gateway relay voice sessions with streaming mic input, realtime audio playback, tool-result bridging, and on-screen transcripts. ([#83130](https://github.com/openclaw/openclaw/issues/83130)) Thanks [@sliekens](https://github.com/sliekens). ##### Fixes - Discord/OpenAI: keep realtime Discord voice sessions hearing follow-up turns with OpenAI realtime and prebuffer assistant playback to avoid choppy starts. ([#80505](https://github.com/openclaw/openclaw/issues/80505)) Thanks [@Solvely-Colin](https://github.com/Solvely-Colin). - Media: prevent image metadata probing from invoking external decoder delegates on unrecognized image bytes, and stop fallback chaining after real processing errors. - Media: install Sharp with the root package and fall back to sips, Windows native imaging, ImageMagick, GraphicsMagick, or ffmpeg for image resizing/conversion when Sharp is unavailable. Fixes [#83401](https://github.com/openclaw/openclaw/issues/83401). Thanks [@scotthuang](https://github.com/scotthuang). - Telegram: deliver generated media completions back into forum topics by preserving topic IDs across requester-agent handoff. ([#83556](https://github.com/openclaw/openclaw/issues/83556)) Thanks [@fuller-stack-dev](https://github.com/fuller-stack-dev). - Gateway: defer update-check startup until after readiness so package update checks no longer block sidecar-ready startup, while preserving update broadcasts and shutdown cleanup. ([#83520](https://github.com/openclaw/openclaw/issues/83520)) Thanks [@samzong](https://github.com/samzong). - Telegram: keep `/btw` and read-only status commands from aborting active runs, and avoid retaining raw update payloads in timed-out spool tombstones. Refs [#83272](https://github.com/openclaw/openclaw/issues/83272). - Agents/video: hide `video_generate` reference-audio parameters unless a registered video provider supports audio inputs. - Plugins/xAI: echo PKCE challenge fields during OAuth authorization-code token exchange for xAI token-endpoint compatibility. ([#83499](https://github.com/openclaw/openclaw/issues/83499)) Thanks [@fuller-stack-dev](https://github.com/fuller-stack-dev). - Codex app-server: hydrate current inbound image attachments before queued runs so Responses-backed agents receive Discord and other channel images as native vision input. Fixes [#83466](https://github.com/openclaw/openclaw/issues/83466). Thanks [@iannwu](https://github.com/iannwu). - Codex app-server: keep native code mode available without forcing code-mode-only so OpenClaw dynamic tool turns complete through the app-server tool bridge. Fixes [#83109](https://github.com/openclaw/openclaw/issues/83109). Thanks [@daswass](https://github.com/daswass). - Release stability: recover stale session diagnostics and Codex OAuth fallback state so stuck runs and reused refresh tokens clear without blocking follow-up work. ([#83503](https://github.com/openclaw/openclaw/issues/83503)) Thanks [@100yenadmin](https://github.com/100yenadmin). - Messages/TTS: apply TTS directives before message-tool sends reach core, gateway, or plugin delivery so opt-in message-tool rooms and proactive sends attach voice notes instead of leaking raw tags. Fixes [#81598](https://github.com/openclaw/openclaw/issues/81598). Thanks [@CG-Intelligence-Agent-Jack](https://github.com/CG-Intelligence-Agent-Jack) and [@CoronovirusG10](https://github.com/CoronovirusG10). - Codex app-server: preserve network access for sandboxed Codex code-mode turns when the OpenClaw sandbox allows outbound egress. Fixes [#83347](https://github.com/openclaw/openclaw/issues/83347). Thanks [@YusukeIt0](https://github.com/YusukeIt0). - QA-Lab: keep the OTLP smoke decoder independent of removed OpenTelemetry generated-root internals. - Messages: default group/channel visible replies to automatic final delivery again, keeping `message_tool` opt-in for ambient/shared rooms and tool-reliable models. - CLI/TUI: force standalone `/exit` runs to terminate after `runTui` returns so onboarding-launched TUI children do not stay alive invisibly. ([#83501](https://github.com/openclaw/openclaw/issues/83501)) Thanks [@fuller-stack-dev](https://github.com/fuller-stack-dev). - Agents/code mode: honor per-agent code-mode config in schema, runtime catalog activation, and model payload filtering. Fixes [#83388](https://github.com/openclaw/openclaw/issues/83388). Thanks [@Kaspre](https://github.com/Kaspre). - Agents/code mode: preserve agent, session, run, and channel context in `before_tool_call` hooks for top-level `exec`/`wait` dispatches. Fixes [#83387](https://github.com/openclaw/openclaw/issues/83387). - QQBot: shorten C2C typing indicators to a 10-second window renewed every 5 seconds, capped to keep a final passive-reply slot available. ([#83469](https://github.com/openclaw/openclaw/issues/83469)) - Replies: keep final payload delivery after live preview updates so channels can finalize or send the completed answer instead of losing preview-only drafts. ([#83468](https://github.com/openclaw/openclaw/issues/83468)) - Discord: deliver final replies in progress-mode preview streams instead of deduplicating the final visible message. ([#83443](https://github.com/openclaw/openclaw/issues/83443)) Thanks [@compoodment](https://github.com/compoodment). - Providers/Xiaomi: replay MiMo Anthropic-compatible `reasoning_content` as provider-required thinking blocks even when OpenClaw thinking is disabled, fixing follow-up tool turns for `mimo-v2-flash`. Fixes [#83407](https://github.com/openclaw/openclaw/issues/83407). Thanks [@Xgenious7](https://github.com/Xgenious7). - Agents/exec approvals: forward approval-runtime credentials on agent-owned Gateway approval calls so approved async commands complete through the existing runtime path instead of stalling on unauthenticated follow-up calls. Thanks [@IWhatsskill](https://github.com/IWhatsskill), [@Patrick-Erichsen](https://github.com/Patrick-Erichsen), and [@jesse-merhi](https://github.com/jesse-merhi). - Gateway/skills: preflight remote macOS skill-bin refreshes with a WebSocket connectivity check so stale node sessions skip quickly instead of logging slow `system.which` timeout warnings. - CLI/config: keep broken discovered plugins that are not referenced by active config from failing `openclaw config validate`, while preserving fatal errors for explicitly configured plugin entries. - GitHub Copilot: drop unsafe native Responses reasoning replay items with non-replayable IDs before dispatch, preventing affected Copilot sessions from failing with `invalid_request_body`. Fixes [#83220](https://github.com/openclaw/openclaw/issues/83220). Thanks [@galiniliev](https://github.com/galiniliev). - Agents/Codex: fail closed when an explicitly requested Codex harness is not registered instead of silently trying configured model fallbacks. Fixes [#83349](https://github.com/openclaw/openclaw/issues/83349). Thanks [@r2-vibes](https://github.com/r2-vibes). - QA-Lab: make runtime tool coverage fail on missing required tool exercise instead of treating pass/pass parity envelope drift as missing coverage. - Core/plugins: harden clawpatch-reported edge cases across gateway auth cleanup, Claude session id paths, plugin activation policy, apply-patch hunk handling, diagnostic redaction, and plugin metadata validation. - UI: show reasoning choices as plain labels instead of leaking internal override wording in session and chat pickers. - Mac app: avoid repeating the Configuration heading inside channel quick settings. - Mac app: keep the Settings sidebar always visible and remove the redundant titlebar hide/show control. - Mac app: prefer explicit private/Tailscale/LAN Gateway endpoints over SSH tunnels, preserve legacy loopback tunnel configs, persist transport choices, and show captured SSH stderr when tunneling really fails. - Gateway/sessions: keep ACP/acpx and runtime child sessions visible in configured-only session lists when their owner or parent session belongs to a configured agent. - Mac app: keep app-level menu commands and Dashboard failure states reachable when the remote Gateway is disconnected. - Mac app: allow longer Gateway and Context errors to wrap in the menu instead of truncating the useful failure detail. - Mac app: tighten remote Gateway fields in Settings so the Connection pane keeps readable labels and full action button text. - Mac app: keep custom Settings card rows left-aligned and full-width so Discovery and status sections no longer appear centered or detached. - Mac app: align Location permission controls to the same trailing column as the rest of Settings. - Mac app: add Dashboard, Chat, Canvas, and Settings shortcuts to the Dock icon menu. - Mac app: replace the Settings window's native split-view sidebar with an explicit layout so page content keeps its leading gutter when the sidebar is shown or hidden. - Mac app: render channel quick config as aligned Settings rows and hide schema-only variants that cannot be edited safely from the quick pane. - Gateway/webchat: hide internal runtime-context and other `display: false` transcript messages from Chat history and live message events. Fixes [#83216](https://github.com/openclaw/openclaw/issues/83216). Thanks [@EmpireCreator](https://github.com/EmpireCreator). - CLI/help: keep `gateway`, `doctor`, `status`, and `health` help registration out of action/runtime imports so subcommand `--help` stays lightweight in constrained terminals. Fixes [#83228](https://github.com/openclaw/openclaw/issues/83228). Thanks [@dfguerrerom](https://github.com/dfguerrerom). - Cron/Discord: keep explicit announce runs in message-tool-only source-reply mode so scheduled agent turns post once instead of also echoing through automatic visible replies. Fixes [#83261](https://github.com/openclaw/openclaw/issues/83261). Thanks [@Theralley](https://github.com/Theralley). - Telegram: preserve forum-topic origin targets in inbound, audio-preflight, and skipped-message hook contexts so follow-up delivery stays bound to the originating topic. Fixes [#83302](https://github.com/openclaw/openclaw/issues/83302). Thanks [@M00zyx](https://github.com/M00zyx). - Telegram: retry HTTP 421 Misdirected Request send failures on a fresh fallback transport so transient edge-node routing errors no longer drop outbound replies. Fixes [#48892](https://github.com/openclaw/openclaw/issues/48892). ([#48908](https://github.com/openclaw/openclaw/issues/48908)) Thanks [@MarsDoge](https://github.com/MarsDoge). - Telegram: fail topic sends closed when Telegram reports `message thread not found` instead of retrying without `message_thread_id` into the base chat. Refs [#83302](https://github.com/openclaw/openclaw/issues/83302). - Config/subagents: remove ignored agent-model `timeoutMs` keys, keep subagent model config to primary/fallback selection, and clean shipped stale config through doctor. Fixes [#83291](https://github.com/openclaw/openclaw/issues/83291). Thanks [@giodl73-repo](https://github.com/giodl73-repo). - Mac app: align the Sessions settings pane with the standard Settings page gutter and row spacing. - OpenAI/Codex: stop rejecting available `openai-codex` GPT-5.1, GPT-5.2, and GPT-5.3 model refs during config validation, while keeping removed Spark aliases suppressed. Fixes [#83303](https://github.com/openclaw/openclaw/issues/83303). - Plugins/xAI: complete OAuth-backed xAI login and sidecar auth fixes, including guarded loopback callback CORS handling, video generation polling/defaults, and native-host User-Agent attribution. ([#83322](https://github.com/openclaw/openclaw/issues/83322)) Thanks [@Jaaneek](https://github.com/Jaaneek). - Codex app-server: preserve streamed native command output in mirrored transcripts and trajectory exports when final snapshots omit aggregated output. ([#83200](https://github.com/openclaw/openclaw/issues/83200)) Thanks [@rozmiarD](https://github.com/rozmiarD). - Codex app-server: fail closed when chat or sender policy denies tools, disabling native code, app, environment, and user MCP surfaces for restricted turns. ([#82374](https://github.com/openclaw/openclaw/issues/82374)) Thanks [@VACInc](https://github.com/VACInc). - Codex app-server: keep recent context-engine messages when oversized projected history is truncated, so short follow-ups in long channel sessions do not fall back to stale earlier turns. ([#83127](https://github.com/openclaw/openclaw/issues/83127)) Thanks [@VACInc](https://github.com/VACInc). - Codex app-server: keep OpenClaw session spawning searchable while steering Codex-native delegation through native subagents, avoiding duplicate direct subagent surfaces. ([#83329](https://github.com/openclaw/openclaw/issues/83329)) Thanks [@fuller-stack-dev](https://github.com/fuller-stack-dev). - Codex app-server: recover stale childless Codex-native subagent task mirrors during maintenance and allow their registry rows to be cancelled without an OpenClaw child session. ([#82836](https://github.com/openclaw/openclaw/issues/82836)) Thanks [@yshimadahrs-ship-it](https://github.com/yshimadahrs-ship-it) and [@joshavant](https://github.com/joshavant). - Feishu: return bound subagent delivery origins from session thread setup so Feishu subagent completions route back to the same DM or topic. ([#83190](https://github.com/openclaw/openclaw/issues/83190)) Thanks [@100menotu001](https://github.com/100menotu001). - CLI/update: tailor post-update Gateway recovery hints by platform, showing systemd, LaunchAgent, Scheduled Task, or generic service-manager guidance instead of macOS-only recovery text. ([#83096](https://github.com/openclaw/openclaw/issues/83096)) Thanks [@rubencu](https://github.com/rubencu). - Plugins: apply a default 15-second timeout to legacy `before_agent_start` hooks so hung plugin handlers no longer block agent startup. Fixes [#48534](https://github.com/openclaw/openclaw/issues/48534). ([#83136](https://github.com/openclaw/openclaw/issues/83136)) Thanks [@therahul-yo](https://github.com/therahul-yo). - Feishu: refresh inbound session delivery context for DM, group, and broadcast turns so later replies do not inherit stale WebChat routing. Fixes [#78274](https://github.com/openclaw/openclaw/issues/78274). - Agents/subagents: require the initial subagent registry save before reporting spawn accepted, returning a spawn error instead of losing an untracked run when the registry write fails. ([#83146](https://github.com/openclaw/openclaw/issues/83146)) Thanks [@yetval](https://github.com/yetval). - QA-Lab/qa-channel: attach redacted agent tool-start traces to outbound `QaBusMessage` records so scenarios can assert actual tool use instead of relying only on reply text. Fixes [#67637](https://github.com/openclaw/openclaw/issues/67637). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: fail live runtime parity reports when assistant-message usage is missing, preventing `0 vs 0` live token rows from being reported as passing proof. Fixes [#80411](https://github.com/openclaw/openclaw/issues/80411). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: add a runtime token-efficiency sidecar report that classifies Codex savings separately from regressions and fails only positive Codex-over-Pi live token deltas above threshold. Fixes [#81093](https://github.com/openclaw/openclaw/issues/81093). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: fail Codex-backed OpenAI live runtime-pair runs before launching isolated workers when no portable Codex auth is available, while staging API-key fallbacks and configured Codex keys for isolated QA agents. Fixes [#80412](https://github.com/openclaw/openclaw/issues/80412). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: refresh parity gates, mock frontier fixtures, model scenarios, and workflow artifact lanes to compare GPT-5.5 against Claude Opus 4.7. Fixes [#74262](https://github.com/openclaw/openclaw/issues/74262). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: make mock parity dispatch provider-aware for source discovery and subagent scenarios so OpenAI and Anthropic lanes no longer share identical canned plans. Fixes [#64879](https://github.com/openclaw/openclaw/issues/64879). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: stop returning Control UI bearer tokens from unauthenticated bootstrap payloads and bind Docker harness ports to loopback-only host addresses. ([#66355](https://github.com/openclaw/openclaw/issues/66355)) Thanks [@pgondhi987](https://github.com/pgondhi987). - Mac app: avoid a SwiftUI metadata crash when rendering the Cron Jobs settings pane. - Agents/subagents: preserve run-mode keep subagent registry entries past the session sweep TTL, so kept subagent runs remain visible after cleanup completes. Fixes [#83132](https://github.com/openclaw/openclaw/issues/83132). ([#83168](https://github.com/openclaw/openclaw/issues/83168)) Thanks [@yetval](https://github.com/yetval). - Agents/OpenAI streams: yield via `setTimeout(0)` instead of `setImmediate` between bursty Responses chunks so abort timers can fire during the yield, keeping cancel-on-timeout responsive on hot streams. Refs [#82462](https://github.com/openclaw/openclaw/issues/82462). - Agents/Codex: keep legacy `oauthRef`-backed OAuth profiles usable while `openclaw doctor --fix` migrates them back to inline credentials, without creating new sidecar credentials. ([#83312](https://github.com/openclaw/openclaw/issues/83312)) Thanks [@joshavant](https://github.com/joshavant). - Agents/Codex: load the selected provider owner alongside the Codex harness runtime so `openai-codex` models resolve when plugin allowlists scope runtime loading. Fixes [#83380](https://github.com/openclaw/openclaw/issues/83380). ([#83519](https://github.com/openclaw/openclaw/issues/83519)) Thanks [@joshavant](https://github.com/joshavant). - Telegram: fail stalled isolated-ingress handlers into tombstones and abort same-lane reply work before restarting, so later same-chat updates drain after a hung turn. Fixes [#83272](https://github.com/openclaw/openclaw/issues/83272). ([#83505](https://github.com/openclaw/openclaw/issues/83505)) Thanks [@joshavant](https://github.com/joshavant). - CLI/config: send SecretRef diagnostics to stderr so JSON command stdout remains parseable. - CLI/doctor: seed Control UI allowed origins when migrating legacy non-loopback gateway bind host aliases like `0.0.0.0`. Fixes [#83286](https://github.com/openclaw/openclaw/issues/83286). Thanks [@giodl73-repo](https://github.com/giodl73-repo). - CLI/plugins: ship the bundled memory CLI as a package entry so package-installed `openclaw memory` commands register correctly. - CLI/update: defer doctor-time plugin package installs during package swaps and seed post-core repair from the updated install registry, preventing duplicate reinstall failures. - CLI/update: preserve old-parent-readable config metadata during legacy package handoffs, fall back only to official `@openclaw/*` npm plugin packages when ClawHub plugin artifacts are unavailable, and keep managed service package roots authoritative during updates. - Feishu: detect SecretRef top-level credentials as a configured default account instead of treating object-backed app secrets as missing. - Gateway/restart: keep ordinary unmanaged SIGUSR1/config restarts in-process instead of detach-spawning an orphaned child, preserving custom supervisor PID tracking while leaving update restarts on the fresh-process path. Fixes [#65668](https://github.com/openclaw/openclaw/issues/65668). - CLI/completion: resolve concrete PowerShell profile paths and reload commands during setup and doctor completion installation. Fixes [#44296](https://github.com/openclaw/openclaw/issues/44296). ([#83059](https://github.com/openclaw/openclaw/issues/83059)) Thanks [@yu-xin-c](https://github.com/yu-xin-c). - Telegram: keep isolated long polling below the hard `getUpdates` request guard so idle bot accounts with high `timeoutSeconds` do not false-disconnect and restart-loop. Fixes [#83264](https://github.com/openclaw/openclaw/issues/83264). Thanks [@riccodecarvalho](https://github.com/riccodecarvalho). - Providers/Google: preserve and recover Gemini 3 tool-call thought signatures during native replay so function-calling turns no longer fail with missing `thought_signature` 400s. Fixes [#72879](https://github.com/openclaw/openclaw/issues/72879). ([#80358](https://github.com/openclaw/openclaw/issues/80358)) Thanks [@abnershang](https://github.com/abnershang). - Telegram: skip transcript-only delivery mirrors and gateway-injected rows when resolving latest assistant text, preventing retained previews from replacing final replies with stale fragments. Fixes [#83159](https://github.com/openclaw/openclaw/issues/83159). ([#83362](https://github.com/openclaw/openclaw/issues/83362)) Thanks [@joshavant](https://github.com/joshavant). - Memory/QMD: keep lexical search on raw hyphenated queries while normalizing semantic QMD sub-searches, avoiding fallback to the builtin index for dashed identifiers and dates. Fixes [#81328](https://github.com/openclaw/openclaw/issues/81328). - Memory-core: distinguish sqlite-vec load failures from missing semantic vector embeddings in degraded `memory index` warnings, so vector recall diagnostics point at unresolved dimensions instead of blaming sqlite-vec when the store is ready. Fixes [#75624](https://github.com/openclaw/openclaw/issues/75624). ([#83056](https://github.com/openclaw/openclaw/issues/83056)) Thanks [@xuruiray](https://github.com/xuruiray) and [@Noah3521](https://github.com/Noah3521). - Agents/subagents: preserve sandbox-peer controller ownership while routing completion announcements back to the originating run session, keeping subagent control and completion delivery scoped correctly. Fixes [#80201](https://github.com/openclaw/openclaw/issues/80201). ([#80242](https://github.com/openclaw/openclaw/issues/80242)) Thanks [@Jerry-Xin](https://github.com/Jerry-Xin). - Gateway: continue restarting remaining channels when one hot-reload channel restart fails, while still reporting aggregate reload failure and rolling back plugin pre-replace stops. Fixes [#83054](https://github.com/openclaw/openclaw/issues/83054). Thanks [@zqchris](https://github.com/zqchris). - Telegram: keep hot-reload restarts from marking polling accounts manually stopped and restart isolated ingress cleanly after worker shutdown, preserving Telegram replies across config reloads. Fixes [#83008](https://github.com/openclaw/openclaw/issues/83008). ([#83410](https://github.com/openclaw/openclaw/issues/83410)) Thanks [@joshavant](https://github.com/joshavant). - Telegram/Ollama: pass current Telegram image attachments into native PI/Ollama vision turns so live photo prompts reach Ollama as native images. Fixes [#83023](https://github.com/openclaw/openclaw/issues/83023). ([#83516](https://github.com/openclaw/openclaw/issues/83516)) Thanks [@joshavant](https://github.com/joshavant). - Gateway/secrets: split the lightweight secrets runtime state and auth-store cache from the full secrets runtime and take a startup fast path when the gateway startup config has no SecretRef values, speeding up secrets startup while preserving cleanup and refresh semantics. - Codex app-server: rotate oversized native Codex threads before resume and cap dynamic tool-result text entering native Codex sessions, preventing stale oversized context from surviving OpenClaw compaction. ([#82981](https://github.com/openclaw/openclaw/issues/82981)) Thanks [@hansolo949](https://github.com/hansolo949). - Gateway/restart: drain pending replies and active chat runs during restart shutdown before sockets and channels close, aborting timed-out chat runs through the normal cleanup path. ([#69121](https://github.com/openclaw/openclaw/issues/69121)) Thanks [@alexlomt](https://github.com/alexlomt). - Agents/Codex: use the Codex runtime context window for OpenAI-model preflight compaction and memory flush checks, so GPT-5.5 Codex sessions compact before hitting the smaller native context limit. Fixes [#82982](https://github.com/openclaw/openclaw/issues/82982). Thanks [@vliuyt](https://github.com/vliuyt). - QA-Lab: clean orphaned gateway temp roots when a suite parent exits and wait on gateway plus transport readiness after config restarts, reducing stale `qa-channel` noise from interrupted runs. Fixes [#65506](https://github.com/openclaw/openclaw/issues/65506). Thanks [@100yenadmin](https://github.com/100yenadmin). - QA-Lab: wake qa-bus long polls that arrive with stale future cursors after a bus restart, preserving reconnect readiness for harness clients. ([#67142](https://github.com/openclaw/openclaw/issues/67142)) Thanks [@hxy91819](https://github.com/hxy91819). - QA-Lab: stage Multipass transfer scripts under OpenClaw's preferred temp root instead of raw OS temp paths, keeping the VM runner inside temp-path guardrails. ([#64098](https://github.com/openclaw/openclaw/issues/64098)) Thanks [@ImLukeF](https://github.com/ImLukeF). - Agents/replies: keep surviving reply media and append a warning when other media references fail, so partial media normalization no longer drops failures silently. Thanks [@Jerry-Xin](https://github.com/Jerry-Xin). - Config/models: accept `thinkingFormat: "together"` in model compat config so Together routes can opt into the Together-specific thinking response shape. - Plugins/tokenjuice: bump the bundled tokenjuice runtime to 0.7.1, bringing Codex hook approval compatibility, pre-tool command wrapping fixes, and Rolldown/Vitest output compaction improvements into the OpenClaw plugin. - Agents/OpenAI: stop post-processing GPT-5 final replies with hardcoded brevity caps, preserving full channel responses instead of appending synthetic ellipses, and log when strict-agentic GPT-5 execution activates. Fixes [#82910](https://github.com/openclaw/openclaw/issues/82910). - Mac app: refine the Settings General and Connection panes with cleaner status panels, card rows, and a single native titlebar sidebar toggle. - Agents/media: deliver failed async image, music, and video generation completions directly when requester-session completion handoff fails, so channel users see provider errors instead of silent fallback stalls. - Browser/CDP: keep loopback proxy bypass active across both `NO_PROXY` casings and redact home-relative Chrome MCP profile paths in attach-failure diagnostics. - Agents/music: steer song, jingle, beat, anthem, and instrumental requests toward `music_generate` audio creation instead of lyric-only replies, and reserve `lyrics` for exact sung words. - Codex app-server: record native Codex tool calls and results into trajectory artifacts so debug/trajectory exports capture the full Codex-native tool history, not just OpenClaw-bridged turns. Thanks [@vyctorbrzezowski](https://github.com/vyctorbrzezowski). - Codex/app-server: keep bound conversation sessions on the owning agent runtime so native Codex control and follow-up turns do not fall back to the default agent client. Fixes [#82954](https://github.com/openclaw/openclaw/issues/82954). ([#82993](https://github.com/openclaw/openclaw/issues/82993)) - CLI/infer: run gateway model probes in fresh explicit sessions so one-shot provider checks do not inherit default agent transcript state. ([#82861](https://github.com/openclaw/openclaw/issues/82861)) Thanks [@Kaspre](https://github.com/Kaspre). - Providers/Together: send video-generation requests to Together's v2 video API even when shared text-model config still points at the v1 base URL. ([#82992](https://github.com/openclaw/openclaw/issues/82992)) - Browser CLI: preserve browser-level options on nested commands, skip option values during lazy command registration, and keep long-running wait/download/dialog hooks open for their advertised wait window. - CLI/sessions: accept `openclaw sessions list` as an alias for `openclaw sessions`, matching other list-style commands. Fixes [#81139](https://github.com/openclaw/openclaw/issues/81139). ([#81163](https://github.com/openclaw/openclaw/issues/81163)) Thanks [@YB0y](https://github.com/YB0y). - Channels/stream previews: widen compact progress draft lines and cut prose at word boundaries while preserving command/path suffixes, with `streaming.progress.maxLineChars` for channel-specific tuning. - CLI/plugins: have `openclaw plugins doctor` warn when a configured runtime needs a missing owner plugin, sharing the same install mapping as `openclaw doctor --fix`. Fixes [#81326](https://github.com/openclaw/openclaw/issues/81326). ([#81674](https://github.com/openclaw/openclaw/issues/81674)) Thanks [@Zavianx](https://github.com/Zavianx). - Agents/Codex: route OpenAI runs that resolve to `openai-codex` through the Codex provider and bootstrap OpenClaw's stored OAuth profile into the Codex harness when the harness owns transport, so `openai/*` model refs no longer fail with `No API key found for openai-codex` despite an existing Codex OAuth profile. ([#82864](https://github.com/openclaw/openclaw/issues/82864)) Thanks [@ragesaq](https://github.com/ragesaq). - Agents/ACP: distinguish prompt-submitted and runtime-active child stalls from true interactive waits, including redacted proxy-env diagnostics for Codex ACP no-output runs. Fixes [#44810](https://github.com/openclaw/openclaw/issues/44810). - Agents/memory: explain that memory-triggered compaction exposes only `read` and append-only `write` when configured core tools are unavailable in `tools.allow` warnings. Fixes [#82941](https://github.com/openclaw/openclaw/issues/82941). Thanks [@galiniliev](https://github.com/galiniliev). - Agents/OpenAI: preserve deterministic tool payload ordering for prompt-cache reuse across OpenAI Responses and chat completions calls. ([#82940](https://github.com/openclaw/openclaw/issues/82940)) Thanks [@galiniliev](https://github.com/galiniliev). - ACP/Codex: honor terminal ACP turn results so failed Codex/acpx runs are not recorded as successful after only progress text. Fixes [#79522](https://github.com/openclaw/openclaw/issues/79522). Thanks [@dudaefj](https://github.com/dudaefj). - Telegram: warn when a media group drops photos that fail to download, including albums where every photo is skipped. Fixes [#55216](https://github.com/openclaw/openclaw/issues/55216). ([#82987](https://github.com/openclaw/openclaw/issues/82987)) Thanks [@eldar702](https://github.com/eldar702). - Agents/skills: apply the full effective tool policy pipeline to inline `command-dispatch: tool` skill dispatch before owner-only filtering, preserving configured allow, deny, sandbox, sender, group, and subagent restrictions. ([#78525](https://github.com/openclaw/openclaw/issues/78525)) - Codex: avoid spawning native hook relay subprocesses for post-tool/finalize events with no registered hook handlers while preserving pre-tool safety and approval relays. Fixes [#76552](https://github.com/openclaw/openclaw/issues/76552). ([#78004](https://github.com/openclaw/openclaw/issues/78004)) Thanks [@evgyur](https://github.com/evgyur). - Channel accounts: keep top-level default channel accounts visible when named accounts are added alongside default credential material, so mixed legacy/new account configs keep resolving `default` instead of silently dropping it. - Agents/CLI: reject empty successful CLI subprocess replies as `empty_response` and keep them out of shared auth-profile health, so blank Claude CLI results no longer become green no-payload turns. Fixes [#83231](https://github.com/openclaw/openclaw/issues/83231). ([#83421](https://github.com/openclaw/openclaw/issues/83421)) Thanks [@joshavant](https://github.com/joshavant). - Codex/Telegram: synthesize native Codex tool progress from final turn snapshots so Telegram `/verbose` stays visible when command events arrive only at completion. - Codex/Telegram: deliver Codex verbose tool summaries in direct message-tool-only turns while suppressing message-send and activity-log noise. ([#83186](https://github.com/openclaw/openclaw/issues/83186)) Thanks [@kurplunkin](https://github.com/kurplunkin). - Mac app: make Channels settings open faster by deferring config-schema work, avoiding startup channel probes, caching decoded channel status rows, and showing only compact quick settings instead of the full generated channel schema. - Control UI: include the Control UI and Gateway protocol versions in protocol-mismatch errors so stale app/dashboard pairings identify which side needs rebuilding or restarting. - Gateway/protocol: restore Gateway WS protocol v4 and keep `message.action` room-event metadata on the existing `inboundTurnKind` wire field while preserving internal inbound-event classification. - Agents/tools: prefer non-webchat session-key routes when the message tool has stale webchat context, so message-tool-only replies keep delivering to the originating channel. Fixes [#82911](https://github.com/openclaw/openclaw/issues/82911). ([#83004](https://github.com/openclaw/openclaw/issues/83004)) Thanks [@joshavant](https://github.com/joshavant). - Channels: keep direct-message last-route writes on isolated `per-channel-peer` sessions instead of contaminating the agent main session with channel delivery context. Fixes [#36614](https://github.com/openclaw/openclaw/issues/36614). Thanks [@aspenas](https://github.com/aspenas). - Mac app: move the Settings sidebar toggle into the native titlebar and tighten the General pane width. - Mac app: keep visited Settings panes mounted so switching tabs no longer blanks and reloads their content. - Mac app: make Config settings open from shallow schema lookups and load selected paths on demand instead of fetching and rendering the full generated config schema up front. - Codex: sanitize inline image payloads before Codex app-server and OpenAI Responses replay, and clear poisoned Codex thread bindings after invalid image errors. Fixes [#82878](https://github.com/openclaw/openclaw/issues/82878). - Providers/GitHub Copilot: request identity-encoded Copilot API responses across token exchange, catalog, model calls, usage, and embeddings so compressed Business-account error payloads no longer reach JSON parsers as gzip bytes. Fixes [#82871](https://github.com/openclaw/openclaw/issues/82871). Thanks [@tonyfe01](https://github.com/tonyfe01). - Telegram: redact nested raw-update identifiers and user metadata before verbose raw update logging, preserving useful update/message ids without exposing chat, user, command, or profile details. ([#82945](https://github.com/openclaw/openclaw/issues/82945)) Thanks [@galiniliev](https://github.com/galiniliev) and [@joshavant](https://github.com/joshavant). - Telegram: preserve replied-to bot messages, captions, and media metadata in group reply chains so follow-up replies understand what the user is reacting to. ([#82863](https://github.com/openclaw/openclaw/issues/82863)) - Providers/Together: update PI runtime packages to 0.74.1 and emit Together-style `reasoning.enabled`/`max_tokens` controls for reasoning-capable OpenAI-completions models. - Agents/diagnostics: split slow embedded-run `attempt-dispatch` startup summaries into workspace, prompt, runtime-plan, and final dispatch subspans so traces identify the delayed setup phase. Fixes [#82782](https://github.com/openclaw/openclaw/issues/82782). ([#82783](https://github.com/openclaw/openclaw/issues/82783)) Thanks [@galiniliev](https://github.com/galiniliev). - Agents/Codex: flatten nested tool-result middleware blocks into bounded text so successful message sends are no longer replaced with `Tool output unavailable due to post-processing error`. Fixes [#82912](https://github.com/openclaw/openclaw/issues/82912). Thanks [@joeykrug](https://github.com/joeykrug). - CLI/media: accept HTTP(S) URLs in `openclaw infer image describe --file`, fetching remote images through the guarded media path instead of treating URLs as local files. Fixes [#82837](https://github.com/openclaw/openclaw/issues/82837). ([#82854](https://github.com/openclaw/openclaw/issues/82854)) Thanks [@neeravmakwana](https://github.com/neeravmakwana). - Agents/subagents: keep session-backed parent runs active when the child wait call times out before the child session has actually settled, so late subagent completions are reconciled instead of being lost. Fixes [#82787](https://github.com/openclaw/openclaw/issues/82787). Thanks [@ramitrkar-hash](https://github.com/ramitrkar-hash). - Control UI: advertise shared Gateway protocol constants in browser connect frames, fixing protocol mismatch handshakes after protocol constant drift. Fixes [#82882](https://github.com/openclaw/openclaw/issues/82882). Thanks [@galiniliev](https://github.com/galiniliev). - Gateway: add rollback protocol-mismatch diagnostics, including client protocol ranges in Gateway logs and deep status/doctor hints for stale client processes. Fixes [#82841](https://github.com/openclaw/openclaw/issues/82841). ([#82908](https://github.com/openclaw/openclaw/issues/82908)) - Agents/subagents: keep successful keep-mode completion payloads pending after final-delivery retry exhaustion, so requester recovery no longer loses final subagent results. Fixes [#82583](https://github.com/openclaw/openclaw/issues/82583). ([#82999](https://github.com/openclaw/openclaw/issues/82999)) Thanks [@joshavant](https://github.com/joshavant). - Gateway/auth: allow same-host trusted-proxy callers to use the documented local direct `gateway.auth.password` fallback after revisiting the [#78684](https://github.com/openclaw/openclaw/issues/78684) fail-closed policy, while keeping token fallback rejected and forwarded-header requests on the trusted-proxy path. Fixes [#82607](https://github.com/openclaw/openclaw/issues/82607). ([#82953](https://github.com/openclaw/openclaw/issues/82953)) Thanks [@joshavant](https://github.com/joshavant). - Agents/subagents: wait for queued completion handoffs to reach the parent transcript before marking them announced, preventing busy parent runs from cleaning up before observing child results. Fixes [#82913](https://github.com/openclaw/openclaw/issues/82913). ([#83039](https://github.com/openclaw/openclaw/issues/83039)) Thanks [@joshavant](https://github.com/joshavant). - Agents/subagents: route group/channel subagent completions through message-tool-only handoffs when required and keep active-requester wake failures from dropping completion delivery. Fixes [#82803](https://github.com/openclaw/openclaw/issues/82803). Thanks [@galiniliev](https://github.com/galiniliev), [@yozakura-ava](https://github.com/yozakura-ava), and [@moeedahmed](https://github.com/moeedahmed). - Memory-core: scan persisted memory source sessions on startup, comparing on-disk transcripts against the index and marking only missing/newer/resized files dirty for incremental sync. Fixes [#82341](https://github.com/openclaw/openclaw/issues/82341). ([#82341](https://github.com/openclaw/openclaw/issues/82341)) Thanks [@giodl73-repo](https://github.com/giodl73-repo). - Telegram: keep the top-level default account in the account list when named accounts or bindings are added alongside top-level credentials, preserving default polling while still letting named-only configs resolve to a single account. Fixes [#82794](https://github.com/openclaw/openclaw/issues/82794). ([#82794](https://github.com/openclaw/openclaw/issues/82794)) Thanks [@giodl73-repo](https://github.com/giodl73-repo). - CLI/models: reuse command-scoped plugin metadata across model listing, provider catalog, auth, and synthetic-auth checks, restoring fast `openclaw models` runs for plugin-heavy installs. Fixes [#82881](https://github.com/openclaw/openclaw/issues/82881). ([#83033](https://github.com/openclaw/openclaw/issues/83033)) Thanks [@joshavant](https://github.com/joshavant). - CLI/channels: show configured official external channels such as Discord in `openclaw channels list` when their plugin package is missing, including the install and doctor repair command instead of reporting no configured channels. Fixes [#82813](https://github.com/openclaw/openclaw/issues/82813). - Signal: preserve mixed-case group IDs through routing and session persistence so group auto-replies keep delivering after updates. Fixes [#82827](https://github.com/openclaw/openclaw/issues/82827). - Agents/tools: keep the `message` tool available in embedded runs when it is explicitly allowed through `tools.alsoAllow` or runtime tool allowlists, so channel plugins with custom reply delivery can still use configured message sends. Fixes [#82833](https://github.com/openclaw/openclaw/issues/82833). Thanks [@cn1313113](https://github.com/cn1313113). - WhatsApp: honor forced document delivery for outbound image, GIF, and video media so `forceDocument`/`asDocument` sends preserve original media bytes instead of using compressed media payloads. ([#79272](https://github.com/openclaw/openclaw/issues/79272)) Thanks [@itsuzef](https://github.com/itsuzef). - WhatsApp: name outbound document attachments from their MIME type when no filename is provided, so PDF and CSV sends arrive as `file.pdf` and `file.csv` instead of an extensionless `file`. Thanks [@mcaxtr](https://github.com/mcaxtr). - Process/diagnostics: report active lane blockers in lane wait warnings so `queueAhead=0` no longer hides commands waiting behind active work. Fixes [#82791](https://github.com/openclaw/openclaw/issues/82791). ([#82792](https://github.com/openclaw/openclaw/issues/82792)) Thanks [@galiniliev](https://github.com/galiniliev). - Process/diagnostics: stop counting the active processing turn as queued backlog in liveness warnings so transient max-only event-loop spikes do not surface as gateway warnings. - Agents/replies: classify provider conversation-state rejections and return a clear message-channel error instead of auto-resetting or falling back to a generic runner failure. ([#82616](https://github.com/openclaw/openclaw/issues/82616)) Thanks [@dutifulbob](https://github.com/dutifulbob). - Browser plugin: trust managed Chrome CDP diagnostics when launch HTTP probes race cold-start readiness, avoiding false startup failures. Fixes [#82904](https://github.com/openclaw/openclaw/issues/82904). ([#82986](https://github.com/openclaw/openclaw/issues/82986)) Thanks [@kmanan](https://github.com/kmanan) and [@hclsys](https://github.com/hclsys). - Android: prompt before replacing a changed Gateway TLS thumbprint, showing the old and new SHA-256 fingerprints so users can accept expected certificate rotations instead of hard failing on pin mismatch. ([#83077](https://github.com/openclaw/openclaw/issues/83077)) Thanks [@sliekens](https://github.com/sliekens). - CLI/status: render extra gateway-like service diagnostics as warning/info output instead of error output. Fixes [#46930](https://github.com/openclaw/openclaw/issues/46930). ([#82922](https://github.com/openclaw/openclaw/issues/82922)) thanks [@giodl73-repo](https://github.com/giodl73-repo). - Agents/failover: classify Moonshot/Kimi exhausted-balance HTTP 429 payloads as billing instead of generic rate limits, preserving billing guidance and fallback behavior. Fixes [#43447](https://github.com/openclaw/openclaw/issues/43447). ([#83079](https://github.com/openclaw/openclaw/issues/83079)) Thanks [@leno23](https://github.com/leno23). - Plugin SDK: bundle `openclaw/plugin-sdk/zod` into the published package artifact and verify the packed zod subpath stays self-contained, so pnpm global installs can register plugins without a package-local `zod` symlink. Fixes [#78398](https://github.com/openclaw/openclaw/issues/78398). ([#78515](https://github.com/openclaw/openclaw/issues/78515)) Thanks [@ggzeng](https://github.com/ggzeng). - Providers/Google: drop compaction-truncated Gemini thought signatures before replay so malformed Base64 no longer aborts the next assistant turn. ([#82995](https://github.com/openclaw/openclaw/issues/82995)) Thanks [@wAngByg](https://github.com/wAngByg). </details> --- ### Configuration 📅 **Schedule**: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined). 🚦 **Automerge**: Disabled by config. Please merge this manually once you are satisfied. ♻ **Rebasing**: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox. 🔕 **Ignore**: Close this PR and you won't be reminded about these updates again. --- - [ ] If you want to rebase/retry this PR, check this box --- This PR has been generated by [Renovate Bot](https://github.com/renovatebot/renovate).  Reviewed-on: https://git.erwanleboucher.dev/eleboucher/homelab/pulls/557

…stic (openclaw#82904) (openclaw#82986) * fix(browser): derive Chrome launch readiness from a single CDP diagnostic (openclaw#82904) The pre-fix launch path used `isChromeReachable` (a lightweight HTTP `/json/version` probe) to decide failure, then called the stronger `diagnoseChromeCdp` only to format the thrown error. On macOS cold starts where the HTTP probe transiently fails *between* the polling loop and the diagnostic call, the runtime would throw "Failed to start Chrome CDP on port ... { ok: true, wsUrl: ... }" — a self-contradicting error containing a successful diagnostic result. Per openclaw#82904 this is the actual user-visible bug. Capture `diagnoseChromeCdp` ONCE after the polling loop and use it for both the decision and the error text. The diagnostic helper already includes the lightweight reachability check and adds a websocket `Browser.getVersion` health command, so it is strictly stronger than the HTTP probe; if `diagnoseChromeCdp` returns ok the launch genuinely succeeded. The existing `withMockChromeCdpServer` success test in chrome.internal.test.ts still exercises this code path end-to-end (real HTTP server + real websocket handshake), so the regression-safety case is covered. The asymmetric `probe-fails-but-diagnostic-succeeds` scenario is hard to mock without restructuring the existing test harness; this commit ships the fix and relies on the upstream ClawSweeper review criteria (manual managed-Chrome cold-start proof) plus the standalone real-behavior probe in the PR body. * fix(browser): import ChromeCdpDiagnostic type from chrome.diagnostics The annotation `let finalDiagnostic: ChromeCdpDiagnostic | null` referenced a type that was only re-exported (not imported) inside chrome.ts, causing oxlint/tsc to read it as the implicit `error` type and fail check-lint, check-prod-types, check-test-types, etc. Add the type to the existing chrome.diagnostics.js import block. * fix(browser): preserve Chrome launch diagnostic fallback * test(browser): satisfy launch diagnostic lint * fix(browser): keep Chrome launch readiness scoped * test(browser): answer CDP launch mock probe --------- Co-authored-by: hclsys <hclsys@users.noreply.github.com> Co-authored-by: Peter Steinberger <steipete@gmail.com>

openclaw-barnacle Bot added size: XS proof: supplied External PR includes structured after-fix real behavior proof. labels May 17, 2026

clawsweeper Bot added P2 Normal backlog priority with limited blast radius. impact:crash-loop Crash, hang, restart loop, or process-level availability failure. labels May 17, 2026

hclsys and others added 5 commits May 17, 2026 10:59

fix(browser): preserve Chrome launch diagnostic fallback

44023d9

test(browser): satisfy launch diagnostic lint

f84def0

fix(browser): keep Chrome launch readiness scoped

26846c9

openclaw-barnacle Bot added size: S and removed size: XS labels May 17, 2026

This comment was marked as low quality.

Sign in to view

test(browser): answer CDP launch mock probe

a762af1

steipete force-pushed the fix-browser-chrome-launch-readiness-from-cdp-diagnostic branch from eee5aef to a762af1 Compare May 17, 2026 10:05

openclaw-barnacle Bot added size: M and removed size: S labels May 17, 2026

steipete merged commit 42435d1 into openclaw:main May 17, 2026
107 of 109 checks passed

steipete mentioned this pull request May 17, 2026

Chrome CDP launch fails despite diagnostic confirming readiness (isChromeReachable vs diagnoseChromeCdp disagreement) #82904

Closed

github-actions Bot mentioned this pull request May 18, 2026

Update ghcr.io/openclaw/openclaw Docker tag to v2026.6.5 - autoclosed claytono/infra#2215

Closed

1 task

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(browser): derive Chrome launch readiness from a single CDP diagnostic (#82904)#82986

fix(browser): derive Chrome launch readiness from a single CDP diagnostic (#82904)#82986
steipete merged 6 commits into
openclaw:mainfrom
hclsys:fix-browser-chrome-launch-readiness-from-cdp-diagnostic

hclsys commented May 17, 2026

Uh oh!

clawsweeper Bot commented May 17, 2026 •

edited

Loading

Uh oh!

This comment was marked as low quality.

steipete commented May 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

hclsys commented May 17, 2026

Problem

Fix

Real behavior proof

Pre-implement audit

Uh oh!

clawsweeper Bot commented May 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment was marked as low quality.

steipete commented May 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

clawsweeper Bot commented May 17, 2026 •

edited

Loading