Skip to content

Fail Codex compaction at the Codex boundary#85958

Merged
pashpashpash merged 7 commits into
mainfrom
codex/codex-compaction-fail-hard
May 25, 2026
Merged

Fail Codex compaction at the Codex boundary#85958
pashpashpash merged 7 commits into
mainfrom
codex/codex-compaction-fail-hard

Conversation

@pashpashpash

@pashpashpash pashpashpash commented May 24, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Summary

Codex runtime compaction was still being treated like an OpenClaw-owned PI safeguard in a few places. OpenClaw could preflight-compact Codex reply turns, post-turn compact Codex CLI transcripts, let a context-engine plugin replace Codex app-server compaction, and force plugin compaction before or after Codex turn/start overflow handling.

That is the wrong boundary. Codex already owns automatic compaction inside the Codex runtime. OpenClaw should not predict when a Codex thread needs compaction and then rewrite the transcript around it. At best that duplicates the native harness. At worst it crosses runtime/auth boundaries, restarts the shared app-server, or drifts a Codex-backed OpenAI session into OpenClaw/public Responses summarization.

This tightens the contract. Automatic compaction for Codex runtime sessions is native Codex-owned. OpenClaw now skips its preflight and CLI automatic compaction paths for Codex, removes the Codex app-server context-engine forced-compaction paths, and no longer lets an ownsCompaction context engine take over Codex app-server compaction. Explicit compaction is still supported: /compact and plugin-requested/manual compaction requests are forwarded to Codex as native thread/compact/start, then OpenClaw returns immediately instead of waiting, timing out, restarting the shared app-server, or falling back to another summarizer.

The fail-closed behavior stays where it belongs. If native Codex compaction cannot be started because the native thread binding is missing or stale, OpenClaw reports that failure honestly. It does not restart the shared Codex app-server as a retry strategy, and it does not fall back to OpenClaw/public OpenAI compaction for Codex runtime sessions. Non-Codex pinned harnesses still keep their previous context-engine fallback path. The Codex harness docs now state that Codex owns native compaction and that context engines do not replace it.

Real behavior proof

Behavior addressed: Codex-runtime sessions should not be automatically compacted by OpenClaw preflight, CLI, context-engine, or public OpenAI summarization paths. Explicit/manual/plugin compaction should only start native Codex compaction through thread/compact/start, without OpenClaw waiting on completion, imposing its own timeout, restarting the shared app-server, counting the pending start as completed OpenClaw compaction, or falling back to another compaction backend.

Real environment tested: Pash's local dev agent was restarted from this branch at head 4b9b8ce0c065ce2e1cda49c027e4f3e2c0e3a295. Gateway reported 2026.5.24 running from /Users/pash/code/openclaw/dist/index.js, pid 56665, healthy loopback on 127.0.0.1:18789. The live proof used the real main agent with Codex OAuth auth redacted, provider openai-codex, model gpt-5.5, and the embedded Codex app-server harness.

Exact steps or command run after this patch: pnpm openclaw gateway restart; pnpm openclaw gateway status --deep; pnpm openclaw agent --agent main --session-key agent:main:codex-compaction-proof-85958-final-1779663763 --message "Final live proof turn 1 for PR 85958 on commit 81328167c4. Reply exactly: FINAL_PROOF_TURN_1_OK" --thinking minimal --timeout 600 --json; pnpm openclaw agent --agent main --session-key agent:main:codex-compaction-proof-85958-final-1779663763 --message "Final live proof turn 2 for PR 85958 on the same session. Reply exactly: FINAL_PROOF_TURN_2_OK" --thinking minimal --timeout 600 --json; after a second gateway restart on the latest head, pnpm openclaw agent --agent main --session-key agent:main:codex-compaction-proof-85958-pending-1779668199 --message "Live pending compact proof setup for PR 85958 on branch head 4b9b8ce0c0. Reply exactly: PENDING_PROOF_SETUP_OK" --thinking minimal --timeout 600 --json; a real Gateway sessions.compact RPC for that scratch session through src/gateway/call.ts; session-store before/after comparison of /Users/pash/.openclaw/agents/main/sessions/sessions.json; a follow-up live turn on the same scratch session after the pending native compact-start signal with pnpm openclaw agent --agent main --session-key agent:main:codex-compaction-proof-85958-pending-1779668199 --message "Post pending native compact-start live proof for PR 85958. Reply exactly: PENDING_PROOF_AFTER_COMPACT_OK" --thinking minimal --timeout 600 --json; rg -n 'thread/compact/start|preflight_compacting|falling back to context engine|Summarization failed|provider_error_4xx|"api":"openai-responses"|"modelApi":"openai-responses"|contextCompaction|gateway_timeout|EMBEDDED FALLBACK' /Users/pash/.openclaw/agents/main/sessions/e15cac51-53a1-440a-b0bf-4200389bfc70*; rg -n 'thread/compact|context_compacted|contextCompaction|Summarization failed|api.responses.write|openai-responses|Incorrect API key|falling back to context engine' /Users/pash/.openclaw/agents/main/agent/codex-home/sessions/2026/05/24/rollout-2026-05-24T17-16-47-019e5c7d-c8de-7570-b60b-96997e8da05d.jsonl; pnpm test src/gateway/server.sessions.compaction.test.ts src/auto-reply/reply/commands-compact.test.ts extensions/codex/src/app-server/compact.test.ts -- --reporter=dot; pnpm exec oxfmt --check --threads=1 extensions/codex/src/app-server/compact.ts extensions/codex/src/app-server/compact.test.ts src/auto-reply/reply/commands-compact.test.ts src/gateway/server.sessions.compaction.test.ts; git diff --check; pnpm check:changed; pnpm build.

Evidence after fix: Normal live turns stayed on the same Codex runtime session. Turn 1 returned exactly FINAL_PROOF_TURN_1_OK with run id 4a70de03-f779-488b-99d4-b2e4e362ea5b, session 7b1d9666-d817-4182-b3a4-43ca985d530f, Codex thread 019e5c3a-2533-7321-845e-44a55ab1545b, provider openai-codex, model gpt-5.5, agentHarnessId: codex, fallbackUsed: false, and runner embedded. Turn 2 returned exactly FINAL_PROOF_TURN_2_OK with run id dfbb3821-57c6-44e7-972e-47e66a71f3ee and the same session/provider/model/harness. The latest-head scratch setup turn returned exactly PENDING_PROOF_SETUP_OK with run id 4a420ac5-9a08-42f8-9054-c8a67adefec9, session e15cac51-53a1-440a-b0bf-4200389bfc70, Codex thread 019e5c7d-c8de-7570-b60b-96997e8da05d, provider openai-codex, model gpt-5.5, agentHarnessId: codex, fallbackUsed: false, runner embedded, and promptTokens: 46541. The real Gateway sessions.compact call returned ok: true, compacted: false, tokensAfter: null, and details { "backend": "codex-app-server", "threadId": "019e5c7d-c8de-7570-b60b-96997e8da05d", "signal": "thread/compact/start", "pending": true }. The session store was unchanged immediately after the pending native compact start: compactionCount remained unset, totalTokens remained 46541, and totalTokensFresh remained true. The Gateway log recorded started codex app-server compaction for the same session and thread. The follow-up turn after that compact-start signal returned exactly PENDING_PROOF_AFTER_COMPACT_OK with run id 3fca6dbc-9bab-4d8d-847b-250138764fda, the same session id, provider openai-codex, model gpt-5.5, agentHarnessId: codex, fallbackUsed: false, and runner embedded. After that normal turn, the session store still had compactionCount unset and updated only normal token bookkeeping to totalTokens: 50857. The Codex rollout recorded the expected native context_compacted event for thread 019e5c7d-c8de-7570-b60b-96997e8da05d, showing Codex completed its own compaction path after OpenClaw started it. The session-file grep found no OpenClaw preflight compaction, context-engine fallback, public openai-responses metadata, Gateway timeout fallback, or embedded fallback artifacts. The Codex rollout grep found no Codex compaction failure, public OpenAI Responses summarization, api.responses.write failure, or context-engine fallback strings.

Observed result after fix: The live dev agent stayed on the Codex runtime for normal turns and explicit compact-start proof, resumed Codex-backed scratch sessions correctly after Gateway restarts, and did not drift into OpenClaw/public Responses compaction. Explicit compaction now behaves as a native Codex start signal: OpenClaw forwards thread/compact/start, reports the start as pending rather than completed, and leaves compaction counters alone while Codex owns the async result internally. The tests lock this down across the app-server, channel /compact, and Gateway sessions.compact seams so future changes do not reintroduce OpenClaw-owned Codex compaction fallbacks or completed-state bookkeeping for pending native starts.

What was not tested: I did not force a destructive live compaction timeout on Pash's running personal dev-agent conversation, and I did not trigger a human-authored Discord or Telegram /compact message from a live chat client. The actual Gateway sessions.compact path was exercised against a real Codex-backed scratch session on the running dev agent, and the channel command behavior is covered by focused tests around the same pending native result shape.

@openclaw-barnacle openclaw-barnacle Bot added agents Agent runtime and tooling extensions: codex size: M maintainer Maintainer-authored PR labels May 24, 2026
@clawsweeper

clawsweeper Bot commented May 24, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Codex review: needs maintainer review before merge. Reviewed May 24, 2026, 11:27 PM ET / 03:27 UTC.

Summary
This PR removes OpenClaw-owned automatic and fallback compaction for Codex runtime sessions, forwards explicit compaction as pending native thread/compact/start, and updates Codex docs, changelog, and focused tests.

PR surface: Source -556, Tests -637, Docs +5. Total -1188 across 18 files.

Reproducibility: yes. Source inspection on current main shows the OpenClaw-owned Codex preflight, context-engine, wait/restart, and fallback compaction paths, and the PR body provides live after-fix Gateway/Codex proof.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🦞 diamond lobster
Patch quality: 🐚 platinum hermit
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:

  • Get maintainer approval on the fail-closed Codex compaction boundary versus the narrower compatibility-preserving path.

Risk before merge

  • Existing Codex runtime users who relied on OpenClaw preflight, CLI, context-engine, public OpenAI summarizer fallback, or app-server restart-on-timeout behavior will now see Codex-native ownership or a visible fail-closed error instead.
  • Missing or stale native Codex thread bindings no longer recover through context-engine compaction, so this can expose session availability failures that the old fallback hid.
  • The related open maintainer PR fix Codex preflight compaction auth routing #86292 is a narrower preflight auth-routing alternative; maintainers should decide whether this broader boundary change supersedes it.
  • This read-only review did not rerun the live Gateway/Codex commands from the PR body.

Maintainer options:

  1. Approve the Codex-owned boundary (recommended)
    If maintainers agree that Codex must own automatic compaction, land this after owner review as an intentional fail-closed behavior change with the existing docs and proof.
  2. Preserve a compatibility mode first
    If existing fallback recovery must remain the default, revise the PR to keep compatibility by default and make strict native-only compaction explicit and tested.
  3. Defer to the narrower auth-routing fix
    If the boundary change is too broad for this release, pause or close this PR and review the narrower open preflight auth-routing PR instead.

Next step before merge
Protected maintainer labeling plus intentional fail-closed compatibility, session-state, and availability changes require maintainer judgment rather than a ClawSweeper repair lane.

Security
Cleared: The diff does not change workflows, dependencies, lockfiles, package scripts, downloaded code, or secret storage; the auth/runtime boundary change is a maintainer compatibility decision rather than a concrete security regression.

Review details

Best possible solution:

Make an explicit maintainer decision on the Codex compaction ownership boundary; if fail-closed native Codex ownership is intended, this PR is the coherent path, otherwise preserve compatibility and land a narrower routing fix such as #86292.

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

Yes. Source inspection on current main shows the OpenClaw-owned Codex preflight, context-engine, wait/restart, and fallback compaction paths, and the PR body provides live after-fix Gateway/Codex proof.

Is this the best way to solve the issue?

Likely yes if maintainers agree with the boundary decision. The patch consistently removes duplicate OpenClaw-owned Codex compaction and aligns docs/tests around native pending starts, but the fail-closed compatibility and availability impact needs owner approval.

Codex review notes: model gpt-5.5, reasoning high; reviewed against 3db1508f1ee7.

Label changes

Label justifications:

  • P1: The PR changes a live Codex agent/channel compaction workflow and can affect real sessions immediately.
  • merge-risk: 🚨 compatibility: The diff intentionally removes fallback compaction and restart recovery paths that existing Codex users may currently depend on.
  • merge-risk: 🚨 session-state: The change alters how pending native compaction updates session counters, token freshness, and transcript ownership around Codex thread state.
  • merge-risk: 🚨 availability: Missing or stale native bindings now fail closed instead of falling back, so compaction attempts can become visible runtime failures.
  • rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🦞 diamond lobster and patch quality is 🐚 platinum hermit.
  • status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes detailed after-fix live Codex/Gateway terminal proof, redacted runtime evidence, negative fallback greps, and focused validation commands for the changed behavior.
  • proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes detailed after-fix live Codex/Gateway terminal proof, redacted runtime evidence, negative fallback greps, and focused validation commands for the changed behavior.
Evidence reviewed

PR surface:

Source -556, Tests -637, Docs +5. Total -1188 across 18 files.

View PR surface stats
Area Files Added Removed Net
Source 7 203 759 -556
Tests 8 431 1068 -637
Docs 3 38 33 +5
Config 0 0 0 0
Generated 0 0 0 0
Other 0 0 0 0
Total 18 672 1860 -1188

What I checked:

  • protected maintainer signal: Live PR metadata lists the protected maintainer label plus compatibility, session-state, and availability merge-risk labels, so this PR should stay open for maintainer handling rather than cleanup closure. (4b9b8ce0c065)
  • current-main Codex compaction behavior: Current main still lets an active context engine own Codex app-server compaction and has native compaction wait/retry/restart bookkeeping around thread/compact/start, which is the behavior this PR intentionally removes. (extensions/codex/src/app-server/compact.ts:51, 3db1508f1ee7)
  • PR native compaction boundary: The PR branch makes maybeCompactCodexAppServerSession only start native Codex compaction, returns immediately, and reports the native start as ok: true, compacted: false, pending: true. (extensions/codex/src/app-server/compact.ts:17, 4b9b8ce0c065)
  • PR automatic preflight skip: The PR branch skips OpenClaw memory-flush preflight compaction when the persisted or policy-resolved runtime is Codex, leaving automatic compaction to the native runtime. (src/auto-reply/reply/agent-runner-memory.ts:633, 4b9b8ce0c065)
  • pending-state coverage: The PR adds Gateway coverage asserting sessions.compact treats Codex native compact start as pending, emits completed: false, and preserves compaction/token counters. (src/gateway/server.sessions.compaction.test.ts:394, 4b9b8ce0c065)
  • real behavior proof: The PR body includes after-fix live dev-agent proof for real Codex-backed turns, a real Gateway sessions.compact call returning pending native start, unchanged session-store counters, a follow-up live turn, negative fallback greps, and focused validation commands. (4b9b8ce0c065)

Likely related people:

  • vincentkoc: Current-main blame for the central Codex compaction and adjacent memory/queued fallback lines points to recent work by Vincent Koc across the affected runtime paths. (role: recent area contributor; confidence: medium; commits: af0776987136; files: extensions/codex/src/app-server/compact.ts, src/auto-reply/reply/agent-runner-memory.ts, src/agents/pi-embedded-runner/compact.queued.ts)
  • steipete: File history for the Codex app-server compaction surface includes the original app-server controls and later stabilization work by Peter Steinberger. (role: feature-history contributor; confidence: medium; commits: 31a0b7bd42a5, 8d72aafdbb8d, 9ac7a0398213; files: extensions/codex/src/app-server/compact.ts, extensions/codex/src/app-server/run-attempt.ts, docs/plugins/codex-harness.md)
  • amknight: The related open member-authored PR addresses Codex preflight compaction auth routing on the overlapping auto-reply memory path, which matters if maintainers choose the narrower compatibility path. (role: adjacent follow-up owner; confidence: medium; commits: 66354a42582b; files: src/auto-reply/reply/agent-runner-memory.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 clawsweeper Bot added rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. labels May 24, 2026
@clawsweeper

clawsweeper Bot commented May 24, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

ClawSweeper PR egg

✨ Hatched: 🥚 common Frosted Clawlet

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: collects tiny proofs.
Image traits: location proof lagoon; accessory miniature diff map; palette cobalt, lime, and pearl; mood sleepy but ready; pose leaning over a miniature review desk; shell starlit enamel shell; lighting golden review-room light; background small review tokens.
Share on X: post this hatch
Copy: My PR egg hatched a 🥚 common Frosted Clawlet 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.

@pashpashpash

pashpashpash commented May 24, 2026

Copy link
Copy Markdown
Contributor Author

@clawsweeper re-review

Addressed the CLI fallback gap: Codex native-harness compaction now fails closed in src/agents/command/cli-compaction.ts, and src/agents/command/cli-compaction.test.ts now asserts stale Codex native bindings do not call context-engine compaction, maintenance, or store recording. The non-Codex pinned-harness fallback test still covers the supported fallback behavior outside Codex.

Validation rerun on the new head: pnpm test extensions/codex/src/app-server/compact.test.ts src/agents/pi-embedded-runner/compact.hooks.test.ts src/agents/command/cli-compaction.test.ts -- --reporter=dot; targeted oxfmt --check; git diff --check; pnpm check:changed; pnpm build.

@clawsweeper

clawsweeper Bot commented May 24, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

🦞🧹
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:

@clawsweeper clawsweeper Bot added P1 High-priority user-facing bug, regression, or broken workflow. merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. merge-risk: 🚨 availability 🚨 May cause crashes, hangs, restart loops, stalls, or process outages. labels May 24, 2026
@pashpashpash pashpashpash force-pushed the codex/codex-compaction-fail-hard branch from c39923d to 83e5f86 Compare May 24, 2026 22:01
@pashpashpash

pashpashpash commented May 24, 2026

Copy link
Copy Markdown
Contributor Author

@clawsweeper re-review

Updated this PR to leave automatic Codex-runtime compaction to Codex itself. OpenClaw now skips reply preflight and CLI automatic compaction for Codex runtime sessions, removes app-server context-engine forced compaction before/after Codex turn overflow handling, and still forwards explicit/manual/plugin compaction requests to native Codex thread/compact/start.

Exact-head Real behavior proof is green on 83e5f860189da1515432b945ff3fcaed1aeceacb. Local validation also passed: pnpm test extensions/codex/src/app-server/compact.test.ts src/auto-reply/reply/agent-runner-memory.test.ts src/agents/command/cli-compaction.test.ts extensions/codex/src/app-server/run-attempt.context-engine.test.ts extensions/codex/src/app-server/run-attempt.test.ts src/agents/pi-embedded-runner/compact.hooks.test.ts -- --reporter=dot; pnpm check:changed; pnpm build; targeted oxfmt --check; git diff --check.

@clawsweeper

clawsweeper Bot commented May 24, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

🦞🧹
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:

@clawsweeper clawsweeper Bot added rating: 🦪 silver shellfish Thin PR readiness signal; proof, validation, or implementation needs work. and removed rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. labels May 24, 2026
@openclaw-barnacle openclaw-barnacle Bot added the commands Command implementations label May 24, 2026
@clawsweeper clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🦐 gold shrimp Decent PR readiness signal, but merge confidence is limited. and removed rating: 🦪 silver shellfish Thin PR readiness signal; proof, validation, or implementation needs work. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. labels May 24, 2026
@clawsweeper

clawsweeper Bot commented May 25, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

🦞🧹
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:

@pashpashpash

pashpashpash commented May 25, 2026

Copy link
Copy Markdown
Contributor Author

@clawsweeper re-review

I added one more live proof point after the pending-state fix: after the real Gateway sessions.compact call returned the pending thread/compact/start result, the same scratch Codex session handled another live turn successfully with fallbackUsed: false, and the Codex rollout recorded the native context_compacted event. The PR body now includes that evidence too.

@clawsweeper

clawsweeper Bot commented May 25, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

🦞👀
ClawSweeper picked this up.

Command router queued. I will update this comment with the next step.

Re-review progress:

@clawsweeper clawsweeper Bot added rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. rating: 🦀 challenger crab Exceptional PR readiness: strong proof, clean patch, and convincing validation. and removed rating: 🦐 gold shrimp Decent PR readiness signal, but merge confidence is limited. status: ⏳ waiting on author ClawSweeper has contributor-facing work open and is waiting for author action. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. rating: 🦀 challenger crab Exceptional PR readiness: strong proof, clean patch, and convincing validation. labels May 25, 2026
@clawsweeper clawsweeper Bot mentioned this pull request May 25, 2026
Closed
@github-actions github-actions Bot mentioned this pull request May 25, 2026
Open
@pashpashpash pashpashpash merged commit dd47e47 into main May 25, 2026
168 of 182 checks passed
@pashpashpash pashpashpash deleted the codex/codex-compaction-fail-hard branch May 25, 2026 05:12
github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request May 25, 2026
@pashpashpash
Fail Codex compaction at the Codex boundary (openclaw#85958)
e85893b
This was referenced May 25, 2026
Closed
Closed
@github-actions github-actions Bot mentioned this pull request May 25, 2026
Open
@pashpashpash pashpashpash mentioned this pull request May 25, 2026
Merged
@masatohoshino masatohoshino mentioned this pull request May 26, 2026
Closed
4 tasks
SebTardif pushed a commit to SebTardif/openclaw that referenced this pull request May 26, 2026
@pashpashpash @SebTardif
Fail Codex compaction at the Codex boundary (openclaw#85958)
f26b387
SebTardif pushed a commit to SebTardif/openclaw that referenced this pull request May 26, 2026
@pashpashpash @SebTardif
Fail Codex compaction at the Codex boundary (openclaw#85958)
c37f30c
SebTardif pushed a commit to SebTardif/openclaw that referenced this pull request May 26, 2026
@pashpashpash @SebTardif
Fail Codex compaction at the Codex boundary (openclaw#85958)
0dd424d
jameslcowan pushed a commit to jameslcowan/openclaw that referenced this pull request Jun 2, 2026
@pashpashpash
Fail Codex compaction at the Codex boundary (openclaw#85958)
d1794a4
SYU8384 pushed a commit to SYU8384/openclaw that referenced this pull request Jun 3, 2026
@pashpashpash
Fail Codex compaction at the Codex boundary (openclaw#85958)
4ff1448
@jalehman jalehman mentioned this pull request Jun 8, 2026
Closed
sablehead pushed a commit to sablehead/openclaw that referenced this pull request Jun 10, 2026
@pashpashpash
Fail Codex compaction at the Codex boundary (openclaw#85958)
10916cb
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

agents Agent runtime and tooling commands Command implementations docs Improvements or additions to documentation extensions: codex gateway Gateway runtime maintainer Maintainer-authored PR merge-risk: 🚨 availability 🚨 May cause crashes, hangs, restart loops, stalls, or process outages. merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. merge-risk: 🚨 session-state 🚨 May lose, corrupt, stale, or mis-associate session, agent, or context state. P1 High-priority user-facing bug, regression, or broken workflow. proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. size: XL status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@pashpashpash