fix(codex): derive terminal-idle watchdog from explicit run timeout by alkor2000 · Pull Request #85296 · openclaw/openclaw

alkor2000 · 2026-05-22T09:38:17Z

Summary

Refs #85242. Derives the Codex app-server terminal-idle watchdog from an explicit per-run timeout instead of always using the hardcoded 30-minute default.

This addresses the early-abort half of #85242 as a focused runtime fix. The diagnostic-wording half (naming the terminal-idle watchdog and its effective timeout in surfaced guidance) remains a separate follow-up, so this PR uses Refs rather than Fixes.

The resolver lives in attempt-timeouts.ts. The call site passes params.runTimeoutOverrideMs, the existing prepared signal for a deliberate cron/CLI timeout, rather than params.timeoutMs, which also carries the implicit 48-hour agent default.

Real behavior proof

Before this patch, a scheduled Codex turn with an explicit timeout above 30 minutes could still be aborted by the 30-minute terminal-idle watchdog.

After the maintainer repair:

no explicit run timeout -> 30-minute terminal-idle floor remains unchanged
explicit 45-minute run timeout -> 45-minute terminal-idle watchdog
explicit 10-minute run timeout -> 30-minute floor remains unchanged
internal harness override -> still wins

Focused proof on the real resolver and run-attempt call path:

node scripts/run-vitest.mjs extensions/codex/src/app-server/attempt-timeouts.test.ts extensions/codex/src/app-server/run-attempt.turn-watches.test.ts -> 72 passed
Blacksmith Testbox tbx_01kwej31e1mn30y1ahjqb23pbz: corepack pnpm check:changed -> passed
oxfmt --check on the three changed files -> clean
fresh autoreview -> no findings

A live 30- or 45-minute scheduled turn was not waited out end to end; timer selection is exercised through the production resolver and the real runCodexAppServerAttempt path. The diagnostic-wording half of #85242 remains intentionally out of scope.

clawsweeper · 2026-05-22T09:40:52Z

Codex review: needs maintainer review before merge. Reviewed June 24, 2026, 3:38 PM ET / 19:38 UTC.

Summary
The PR derives the Codex app-server terminal-idle watchdog from the effective run timeout, passes that budget from run-attempt, and adds focused resolver tests.

PR surface: Source +11, Tests +19. Total +30 across 3 files.

Reproducibility: yes. at source level: current main and v2026.6.10 resolve terminal-idle from the 30-minute constant and do not pass params.timeoutMs. I did not live-wait for a 30-minute scheduled Codex turn.

Review metrics: 1 noteworthy metric.

Internal timeout default: 1 changed. The terminal-idle watchdog now follows the existing effective run budget when no explicit override is set, so maintainers should notice the runtime behavior change before merge.

Root-cause cluster
Relationship: fixed_by_candidate
Canonical: #85242
Summary: This PR is the open candidate for the runtime-budget half of the canonical Codex terminal-idle timeout issue; diagnostic wording remains tracked there.

Members:

canonical: Codex app-server terminal idle timeout is hardcoded and reports misleading config guidance #85242 - The issue explicitly tracks both the hardcoded terminal-idle budget and the misleading timeout guidance.
superseded: fix(codex): align terminal idle timeout with run budget #90926 - The closed unmerged PR attempted a broader runtime-plus-diagnostics fix, but it is not a viable replacement for this open branch.
partial_overlap: v2026.5.20 still requires local stability hotfixes after auto-update #85852 - The umbrella release-stability report includes this terminal-idle timeout problem among several local hotfixes.
partial_overlap: [Bug]: Codex app-server session wedges indefinitely after a 429 rate-limit burst — terminal-idle watchdog still gated by activeAppServerTurnRequests > 0 (regression/incomplete fix of #82681) #89742 - The issue shares the terminal-idle watchdog area but reports a separate active-request-gated wedge after rate-limit bursts.

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🐚 platinum hermit
Patch quality: 🦞 diamond lobster
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

none.

Risk before merge

[P1] The terminal-idle-specific diagnostic wording remains intentionally out of scope; Codex app-server terminal idle timeout is hardcoded and reports misleading config guidance #85242 should stay open if this runtime fix lands.
[P1] The proof exercises the real resolver under Node rather than a live 30-plus-minute scheduled Codex turn; that is acceptable for this narrow resolver change but not full end-to-end runtime proof.

Maintainer options:

Decide the mitigation before merge
Land the focused runtime-budget fix if maintainers accept the split scope, and keep Codex app-server terminal idle timeout is hardcoded and reports misleading config guidance #85242 open for terminal-idle-specific timeout guidance.
Pause or close
Do not merge this PR until maintainers decide whether the risk is worth taking.

Next step before merge

No automated repair is indicated; maintainers should decide whether to merge this runtime-only fix while preserving the diagnostics follow-up.

Security
Cleared: The diff only changes internal Codex timeout resolution, one call site, and tests; it adds no dependency, workflow, permission, secret, package, or generated-code surface.

Review details

Best possible solution:

Land the focused runtime-budget fix if maintainers accept the split scope, and keep #85242 open for terminal-idle-specific timeout guidance.

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

Yes at source level: current main and v2026.6.10 resolve terminal-idle from the 30-minute constant and do not pass params.timeoutMs. I did not live-wait for a 30-minute scheduled Codex turn.

Is this the best way to solve the issue?

Yes for the early-abort half: deriving the wrapper-owned watchdog from the effective run budget is narrower than adding a new config knob or upstream protocol field. The terminal-idle guidance remains a separate follow-up.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against fa6a9509bc95.

Label changes

Label justifications:

P2: This is a normal-priority Codex runtime bug fix for long scheduled app-server turns being cut short by an internal terminal-idle watchdog.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🦞 diamond lobster.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body includes after-fix terminal output from invoking the real resolver under Node, showing a 45-minute budget resolves to 2700000 ms while defaults and overrides are preserved.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes after-fix terminal output from invoking the real resolver under Node, showing a 45-minute budget resolves to 2700000 ms while defaults and overrides are preserved.

Evidence reviewed

PR surface:

Source +11, Tests +19. Total +30 across 3 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	2	13	2	+11
Tests	1	19	0	+19
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	3	32	2	+30

What I checked:

Root policy read: Root AGENTS.md was read fully and its Codex dependency-inspection, scoped AGENTS, whole-path review, and ClawSweeper PR review rules apply to this Codex runtime PR. (AGENTS.md:1, fa6a9509bc95)
Scoped extension policy read: The bundled plugin boundary guide was read because the changed Codex runtime lives under extensions/codex; the PR stays inside the Codex plugin boundary and does not expand plugin SDK/API surface. (extensions/AGENTS.md:1, fa6a9509bc95)
Current main source behavior: Current main still resolves terminal idle from only the explicit override and CODEX_TURN_TERMINAL_IDLE_TIMEOUT_MS, so the effective run budget is not considered. (extensions/codex/src/app-server/attempt-timeouts.ts:120, fa6a9509bc95)
Current main caller behavior: runCodexAppServerAttempt currently calls resolveCodexTurnTerminalIdleTimeoutMs with options.turnTerminalIdleTimeoutMs only, while params.timeoutMs is available nearby as the effective run budget. (extensions/codex/src/app-server/run-attempt.ts:1573, fa6a9509bc95)
PR head resolver behavior: PR head adds effectiveRunTimeoutMs, normalizes it through the existing helper, keeps the 30-minute floor, and lets longer run budgets raise the default while preserving explicit override precedence. (extensions/codex/src/app-server/attempt-timeouts.ts:120, 715976872f50)
PR head call site: PR head passes params.timeoutMs into the terminal-idle resolver from runCodexAppServerAttempt, connecting the existing effective run timeout to the watchdog. (extensions/codex/src/app-server/run-attempt.ts:1564, 715976872f50)

Likely related people:

steipete: Recent history ties this account to the app-server attempt split and idle timer caps in the affected timeout/run-attempt/watch area, and PR discussion records the split-scope review note. (role: recent area contributor and reviewer; confidence: high; commits: a4c2e7f5cf1b, 62abfd3dcb9a, 4135771adfca; files: extensions/codex/src/app-server/attempt-timeouts.ts, extensions/codex/src/app-server/run-attempt.ts, extensions/codex/src/app-server/attempt-turn-watches.ts)
kevinslin: Recent merged work added Codex app-server completion-timeout diagnostics in the same result, watch, and attempt surfaces that own the remaining terminal-idle guidance follow-up. (role: adjacent diagnostic contributor; confidence: medium; commits: ab7c92282522; files: extensions/codex/src/app-server/attempt-results.ts, extensions/codex/src/app-server/attempt-turn-watches.ts, extensions/codex/src/app-server/run-attempt.ts)
fuller-stack-dev: Recent merged history preserved Codex post-tool reasoning liveness in attempt-timeouts.ts and adjacent turn-watch tests, close to this watchdog behavior. (role: recent liveness-path contributor; confidence: medium; commits: cf378e4cc8c4, 77a682c5def2; files: extensions/codex/src/app-server/attempt-timeouts.ts, extensions/codex/src/app-server/run-attempt.turn-watches.test.ts, extensions/codex/src/app-server/attempt-results.ts)
vincentkoc: Recent current-main history touches run-attempt.ts heavily across Codex runtime behavior and prompt/lifecycle changes, making this account a useful routing candidate for merge coordination. (role: recent area contributor; confidence: medium; commits: 258b83c4386c, 9f675920bf5a, 2b75806197ab; files: extensions/codex/src/app-server/run-attempt.ts)

What the crustacean ranks mean

🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

clawsweeper · 2026-05-22T09:46:33Z

ClawSweeper PR egg

✨ Hatched: 🥚 common Neon Crabkin

Hatch command

Comment @clawsweeper hatch when this PR is hatchable.

Hatchability rules:

Merged PRs are hatchable.
Open PRs are hatchable when they are status: 👀 ready for maintainer look, status: 🚀 automerge armed, or labeled clawsweeper:automerge.
Closed unmerged PRs are hatchable only when one of those hatchable labels is still present in the durable record.

Rarity: 🥚 common.
Trait: stacks clean commits.
Image traits: location green-check meadow; accessory release bell; palette moss green and polished brass; mood celebratory; pose leaning over a miniature review desk; shell frosted glass shell; lighting clean product lighting; background delicate sparkle particles.
Share on X: post this hatch
Copy: My PR egg hatched a 🥚 common Neon Crabkin in ClawSweeper.

What is this egg doing here?

Eggs appear after the PR passes real-behavior proof. It is here for vibes, not verdicts: it does not change labels, ratings, merge decisions, or automation.
The shell reacts to review momentum: open follow-up work warms it up, re-review makes it wobble, and a clean final review lets it hatch.
Hatchability usually comes from sufficient real-behavior proof, no blocking P0/P1/P2 findings, no security attention needed, and clean correctness. A merged PR is already final, so merge makes the egg hatchable independently.
The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.

ai-hpc · 2026-05-22T13:41:02Z

This looks like the right narrow fix for the early-abort path: deriving the terminal-idle watchdog from the effective run timeout avoids cutting long scheduled Codex turns back to the internal 30m default.

One remaining issue-scope question: #85242 also asks for terminal-idle-specific user guidance. If this PR is meant to fully close the issue, I think the timeout message should name the terminal-idle watchdog and the effective timeout that fired. Otherwise this can land as the runtime fix, with diagnostics tracked separately.

steipete · 2026-05-22T13:44:46Z

Reviewed the diff and I agree with @ai-hpc's scope note: the runtime fix itself looks narrow and correct, but this should not fully close #85242 as-is.

The PR derives turnTerminalIdleTimeoutMs from params.timeoutMs and covers the resolver behavior with focused tests, so the 30m early-abort path is addressed. The missing part is the second half of the issue: if this watchdog still fires, the surfaced guidance should identify the terminal-idle watchdog/effective timeout instead of only telling users to increase agents.defaults.timeoutSeconds.

So I would either:

add the diagnostic wording update here, then keep Fixes #85242; or
land this as the runtime fix and change the closure language to leave the diagnostics follow-up tracked separately.

alkor2000 · 2026-05-30T17:28:41Z

Thanks @ai-hpc @steipete for the review. Going with the second option: landing this as the focused runtime fix and tracking the diagnostics wording as a separate follow-up.

Changes since the review:

Rebased onto current main. The resolver moved to attempt-timeouts.ts in the app-server split, so the derivation now lives there, reusing the existing resolvePositiveIntegerTimeoutMs helper to match the neighbouring post-tool resolver. The call site in run-attempt.ts passes params.timeoutMs.
Changed the closure language from Fixes #85242 to Refs #85242 so the issue stays open for the diagnostics follow-up (surfacing the terminal-idle watchdog and effective timeout in the guidance, rather than only pointing at agents.defaults.timeoutSeconds).
Re-ran the real-resolver proof under node and added the budget-derivation unit tests in attempt-timeouts.test.ts.

Happy to open the diagnostics follow-up separately, or fold it in here if you'd prefer a single PR after all.

alkor2000 · 2026-05-30T17:54:50Z

@clawsweeper re-review

The three failing checks are unrelated to this diff (which only touches extensions/codex/src/app-server: attempt-timeouts.ts, run-attempt.ts, and the colocated test). They are flaky/long-running shards in unrelated subsystems:

checks-node-agentic-agents-core: fails in src/agents/code-mode.test.ts > 'terminates hostile infinite loops outside the main event loop' (a timing-sensitive loop-termination test).
checks-node-agentic-commands-doctor: the shard ran ~684s before failing, in doctor config-validation tests.
checks-node-core-runtime-infra-state: a long-running infra-state shard.

None touch the Codex app-server timeout code. Locally the focused suites pass (attempt-timeouts.test.ts and run-attempt.test.ts, Test Files 2 passed). A CI re-run should clear these.

clawsweeper · 2026-05-30T17:54:52Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26690877074
Updated: 2026-05-30T18:02:06.563Z

Fixes the early-abort half of openclaw#85242. The Codex app-server terminal-idle watchdog used a hardcoded 30-minute default that was not derived from the effective run timeout, so a scheduled turn configured with a longer timeoutSeconds could be aborted early at 30 minutes even with budget left. resolveCodexTurnTerminalIdleTimeoutMs (now in attempt-timeouts.ts after the upstream split) accepts the effective run timeout and, with no explicit override, follows the run budget instead of the 30-minute default: - explicit override always wins (advanced config / tests) - otherwise terminal-idle = max(30min floor, run budget), so a longer run is no longer cut short and existing protection is never shortened - falls back to the 30min default when no run budget is known Reuses the existing resolvePositiveIntegerTimeoutMs helper, matching the neighbouring post-tool resolver. Adds focused unit tests for the derivation. The diagnostic-wording half of openclaw#85242 (naming the terminal-idle watchdog in the surfaced guidance) is left as a separate follow-up.

vincentkoc · 2026-07-01T10:24:55Z

Land-ready at 0cc6afb76909565103e85b520faff92cbff3c018.

The reported bug is valid: an explicitly configured run timeout above 30 minutes could still be cut off by the Codex terminal-idle watchdog at 30 minutes. During review, the original patch was repaired so the watchdog follows only the explicit run-timeout override. Normal runs retain the 30-minute liveness guard, and explicit no-timeout remains bounded by the existing maximum idle watchdog.

Evidence:

node scripts/run-vitest.mjs extensions/codex/src/app-server/attempt-timeouts.test.ts extensions/codex/src/app-server/run-attempt.turn-watches.test.ts - 72 passed
Blacksmith Testbox tbx_01kwej31e1mn30y1ahjqb23pbz - corepack pnpm check:changed passed
fresh autoreview - 0 accepted/actionable findings
exact-head CI run 28509879614 - passed after rerunning one unrelated compact-node shard that had timed out in src/agents/tools/message-tool.test.ts
native scripts/pr prepare-run 85296 - passed with exact-head hosted gates

Direct Codex contract check:

../codex/codex-rs/app-server-protocol/src/protocol/v2/turn.rs:68 - turn start has no timeout/deadline field
../codex/codex-rs/app-server-protocol/src/protocol/v2/turn.rs:208 - interruption is a separate thread/turn operation
../codex/codex-rs/app-server/src/request_processors/turn_processor.rs:157 - app-server starts the turn without a timeout contract

Issue #85242 remains open because this PR fixes timeout behavior, not the broader diagnostic wording request.

vincentkoc · 2026-07-01T10:25:18Z

Merged via squash.

Prepared head SHA: 0cc6afb76909565103e85b520faff92cbff3c018
Landed commit: 4076ba0bd9e89d7a2056d3b5c0c3b8f8689cb9ea

…penclaw#85296) * fix(codex): derive terminal-idle watchdog from effective run timeout Fixes the early-abort half of openclaw#85242. The Codex app-server terminal-idle watchdog used a hardcoded 30-minute default that was not derived from the effective run timeout, so a scheduled turn configured with a longer timeoutSeconds could be aborted early at 30 minutes even with budget left. resolveCodexTurnTerminalIdleTimeoutMs (now in attempt-timeouts.ts after the upstream split) accepts the effective run timeout and, with no explicit override, follows the run budget instead of the 30-minute default: - explicit override always wins (advanced config / tests) - otherwise terminal-idle = max(30min floor, run budget), so a longer run is no longer cut short and existing protection is never shortened - falls back to the 30min default when no run budget is known Reuses the existing resolvePositiveIntegerTimeoutMs helper, matching the neighbouring post-tool resolver. Adds focused unit tests for the derivation. The diagnostic-wording half of openclaw#85242 (naming the terminal-idle watchdog in the surfaced guidance) is left as a separate follow-up. * fix(codex): derive terminal-idle watchdog from effective run timeout * fix(codex): preserve default terminal idle watchdog --------- Co-authored-by: openclaw-clownfish[bot] <280122609+openclaw-clownfish[bot]@users.noreply.github.com> Co-authored-by: Vincent Koc <vincentkoc@ieee.org>

…penclaw#85296) * fix(codex): derive terminal-idle watchdog from effective run timeout Fixes the early-abort half of openclaw#85242. The Codex app-server terminal-idle watchdog used a hardcoded 30-minute default that was not derived from the effective run timeout, so a scheduled turn configured with a longer timeoutSeconds could be aborted early at 30 minutes even with budget left. resolveCodexTurnTerminalIdleTimeoutMs (now in attempt-timeouts.ts after the upstream split) accepts the effective run timeout and, with no explicit override, follows the run budget instead of the 30-minute default: - explicit override always wins (advanced config / tests) - otherwise terminal-idle = max(30min floor, run budget), so a longer run is no longer cut short and existing protection is never shortened - falls back to the 30min default when no run budget is known Reuses the existing resolvePositiveIntegerTimeoutMs helper, matching the neighbouring post-tool resolver. Adds focused unit tests for the derivation. The diagnostic-wording half of openclaw#85242 (naming the terminal-idle watchdog in the surfaced guidance) is left as a separate follow-up. * fix(codex): derive terminal-idle watchdog from effective run timeout * fix(codex): preserve default terminal idle watchdog --------- Co-authored-by: openclaw-clownfish[bot] <280122609+openclaw-clownfish[bot]@users.noreply.github.com> Co-authored-by: Vincent Koc <vincentkoc@ieee.org> (cherry picked from commit 4076ba0)

…s not persisted The fork's fallback-no-persist gate (b8e4198) wrapped the whole finalizeCronRun persistence block, so a fallback run also skipped clearCliSession/setCliSessionId maintenance for the CLI provider that actually ran, leaving stale bindings to hijack the next CLI run. Upstream runs binding maintenance unconditionally (only aborted runs skip it); restore that shape and keep the no-persist gate narrowly around setSessionRuntimeModel/contextTokens. Fixes the base regression test 'clears stale CLI bindings when cron CLI replacement is unflushed' (failing on the branch since b8e4198; bisected). Also: repoint the openclaw#98356 test's getCliSessionBinding import to the 6.11 home (src/agents/cli-session.ts), fix three oxlint errors (promise-executor style in the openclaw#85296 test hunk; only-throw-error + no-shadow in the memory-lancedb failover port).

…penclaw#85296) * fix(codex): derive terminal-idle watchdog from effective run timeout Fixes the early-abort half of openclaw#85242. The Codex app-server terminal-idle watchdog used a hardcoded 30-minute default that was not derived from the effective run timeout, so a scheduled turn configured with a longer timeoutSeconds could be aborted early at 30 minutes even with budget left. resolveCodexTurnTerminalIdleTimeoutMs (now in attempt-timeouts.ts after the upstream split) accepts the effective run timeout and, with no explicit override, follows the run budget instead of the 30-minute default: - explicit override always wins (advanced config / tests) - otherwise terminal-idle = max(30min floor, run budget), so a longer run is no longer cut short and existing protection is never shortened - falls back to the 30min default when no run budget is known Reuses the existing resolvePositiveIntegerTimeoutMs helper, matching the neighbouring post-tool resolver. Adds focused unit tests for the derivation. The diagnostic-wording half of openclaw#85242 (naming the terminal-idle watchdog in the surfaced guidance) is left as a separate follow-up. * fix(codex): derive terminal-idle watchdog from effective run timeout * fix(codex): preserve default terminal idle watchdog --------- Co-authored-by: openclaw-clownfish[bot] <280122609+openclaw-clownfish[bot]@users.noreply.github.com> Co-authored-by: Vincent Koc <vincentkoc@ieee.org> (cherry picked from commit 4076ba0)

…s not persisted The fork's fallback-no-persist gate (b8e4198) wrapped the whole finalizeCronRun persistence block, so a fallback run also skipped clearCliSession/setCliSessionId maintenance for the CLI provider that actually ran, leaving stale bindings to hijack the next CLI run. Upstream runs binding maintenance unconditionally (only aborted runs skip it); restore that shape and keep the no-persist gate narrowly around setSessionRuntimeModel/contextTokens. Fixes the base regression test 'clears stale CLI bindings when cron CLI replacement is unflushed' (failing on the branch since b8e4198; bisected). Also: repoint the openclaw#98356 test's getCliSessionBinding import to the 6.11 home (src/agents/cli-session.ts), fix three oxlint errors (promise-executor style in the openclaw#85296 test hunk; only-throw-error + no-shadow in the memory-lancedb failover port). (cherry picked from commit a01b392)

openclaw-barnacle Bot added extensions: codex size: S proof: supplied External PR includes structured after-fix real behavior proof. labels May 22, 2026

alkor2000 force-pushed the fix-85242-codex-terminal-idle-timeout branch from 6ebefb8 to 1f23f29 Compare May 30, 2026 17:28

openclaw-barnacle Bot added size: XS and removed size: S labels May 30, 2026

clawsweeper Bot added rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. and removed rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. labels Jun 20, 2026

vincentkoc force-pushed the fix-85242-codex-terminal-idle-timeout branch from 1f23f29 to 0408304 Compare June 23, 2026 01:44

alkor2000 and others added 3 commits July 1, 2026 03:07

fix(codex): derive terminal-idle watchdog from effective run timeout

60f6637

fix(codex): preserve default terminal idle watchdog

0cc6afb

vincentkoc force-pushed the fix-85242-codex-terminal-idle-timeout branch from 7159768 to 0cc6afb Compare July 1, 2026 10:08

openclaw-barnacle Bot added size: S and removed size: XS labels Jul 1, 2026

vincentkoc changed the title ~~fix(codex): derive terminal-idle watchdog from effective run timeout~~ fix(codex): derive terminal-idle watchdog from explicit run timeout Jul 1, 2026

openclaw-barnacle Bot added the triage: needs-pr-context Candidate: external PR body lacks required problem context or evidence. label Jul 1, 2026

vincentkoc self-assigned this Jul 1, 2026

vincentkoc merged commit 4076ba0 into openclaw:main Jul 1, 2026
159 of 166 checks passed

Navigation Menu

Uh oh!

fix(codex): derive terminal-idle watchdog from explicit run timeout - #85296

fix(codex): derive terminal-idle watchdog from explicit run timeout#85296
vincentkoc merged 3 commits into
openclaw:mainfrom
alkor2000:fix-85242-codex-terminal-idle-timeout

alkor2000 commented May 22, 2026 •

edited by vincentkoc

Loading

Uh oh!

clawsweeper Bot commented May 22, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 22, 2026

Uh oh!

ai-hpc commented May 22, 2026

Uh oh!

steipete commented May 22, 2026

Uh oh!

alkor2000 commented May 30, 2026

Uh oh!

alkor2000 commented May 30, 2026

Uh oh!

clawsweeper Bot commented May 30, 2026 •

edited

Loading

Uh oh!

vincentkoc commented Jul 1, 2026

Uh oh!

Uh oh!

vincentkoc commented Jul 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

alkor2000 commented May 22, 2026 • edited by vincentkoc Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Real behavior proof

Uh oh!

clawsweeper Bot commented May 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 22, 2026

Hatch command

Uh oh!

ai-hpc commented May 22, 2026

Uh oh!

steipete commented May 22, 2026

Uh oh!

alkor2000 commented May 30, 2026

Uh oh!

alkor2000 commented May 30, 2026

Uh oh!

clawsweeper Bot commented May 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vincentkoc commented Jul 1, 2026

Uh oh!

Uh oh!

vincentkoc commented Jul 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

alkor2000 commented May 22, 2026 •

edited by vincentkoc

Loading

clawsweeper Bot commented May 22, 2026 •

edited

Loading

clawsweeper Bot commented May 30, 2026 •

edited

Loading