fix(openrouter): use endpoint context limits by clawsweeper[bot] · Pull Request #86041 · openclaw/openclaw

clawsweeper · 2026-05-24T12:31:21Z

Makes #85949 merge-ready for the ClawSweeper automerge loop.
The edit pass should inspect the live PR diff, review comments, and failing checks; rebase if needed; keep the contributor branch credited; and stop only when validation is green or an external blocker is proven.

ClawSweeper 🐠 replacement reef notes:

Cluster: automerge-openclaw-openclaw-85949
Source PRs: fix(openrouter): use endpoint context limits #85949
Credit: Source PR: fix(openrouter): use endpoint context limits #85949
Validation: pnpm check:changed
Replacement reason: ClawSweeper could not update the source PR branch directly, so it opened a writable replacement PR instead.
Automerge requested by: @Takhoffman

Repair fallback: GitHub rejected the repair branch push because it updates workflow files and the ClawSweeper app token does not have workflows permission

Inherited issue-closing references from the source PR:
Fixes #85921

Co-author credit kept:

@TurboTheTurtle: Co-authored-by: Andy Ye 35905412+TurboTheTurtle@users.noreply.github.com

fish notes: model gpt-5.5, reasoning high; reviewed against d193cae.

clawsweeper · 2026-05-24T12:33:24Z

Codex review: passed.

Latest ClawSweeper review: 2026-05-24 13:28 UTC / May 24, 2026, 9:28 AM ET.

Workflow note: Future ClawSweeper reviews update this same comment in place.

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.

PR Surface
Source +1, Tests +106, Docs +1, Other +49. Total +157 across 6 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	1	3	2	+1
Tests	1	107	1	+106
Docs	1	2	1	+1
Config	0	0	0	0
Generated	0	0	0	0
Other	3	49	0	+49
Total	6	161	4	+157

Summary
The branch updates OpenRouter dynamic model capability parsing to prefer top_provider.context_length, bumps the disk cache version, adds regression coverage and a changelog entry, and adds script helper declaration files.

Reproducibility: yes. from source and live catalog evidence rather than an authenticated inference turn. Current main reads the top-level context_length, and the public OpenRouter catalog currently reports a smaller endpoint-specific top_provider.context_length for the reported model.

PR rating
Overall: 🦞 diamond lobster
Proof: 🦞 diamond lobster
Patch quality: 🦞 diamond lobster
Summary: Strong PR: live catalog proof and focused regression tests support a small provider-parser fix; only normal exact-head CI completion remains.

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.

Real behavior proof
Sufficient (live_output): The source PR body includes live public OpenRouter catalog output and parsed after-fix capability output showing both contextWindow and maxTokens resolving to endpoint values.

Risk before merge

Two GitHub checks for the exact PR head were still in progress during this review, so merge should wait for the normal exact-head check gate.
Authenticated OpenRouter inference was not run; the available proof covers the live public catalog shape, parser output, and targeted model/transport validation.

Maintainer options:

Decide the mitigation before merge
Land the narrow parser and cache-version fix after exact-head checks complete so the linked OpenRouter context-overflow regression can close.
Pause or close
Do not merge this PR until maintainers decide whether the risk is worth taking.

Next step before merge
No repair job is needed; the current PR head already carries the narrow fix and should proceed through exact-head CI and automerge gates.

Security
Cleared: The diff changes provider parser logic, tests, changelog text, and type-only script declarations; it does not add dependencies, workflows, lockfile changes, secrets handling, or new code-execution paths.

Review details

Best possible solution:

Land the narrow parser and cache-version fix after exact-head checks complete so the linked OpenRouter context-overflow regression can close.

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

Yes, from source and live catalog evidence rather than an authenticated inference turn. Current main reads the top-level context_length, and the public OpenRouter catalog currently reports a smaller endpoint-specific top_provider.context_length for the reported model.

Is this the best way to solve the issue?

Yes. Preferring the endpoint-specific limit while keeping the top-level fallback and invalidating old disk caches is the narrow provider-owned fix.

Label justifications:

P1: The linked regression can make OpenRouter agent turns repeatedly fail with context-overflow errors for real users.
rating: 🦞 diamond lobster: Current PR rating is 🦞 diamond lobster because proof is 🦞 diamond lobster, patch quality is 🦞 diamond lobster, and Strong PR: live catalog proof and focused regression tests support a small provider-parser fix; only normal exact-head CI completion remains.
status: 🚀 automerge armed: This PR is in ClawSweeper's automerge lane. Sufficient (live_output): The source PR body includes live public OpenRouter catalog output and parsed after-fix capability output showing both contextWindow and maxTokens resolving to endpoint values.
proof: sufficient: Contributor real behavior proof is sufficient. The source PR body includes live public OpenRouter catalog output and parsed after-fix capability output showing both contextWindow and maxTokens resolving to endpoint values.

What I checked:

Current main still has the old parser behavior: parseModel in current main sets contextWindow from model.context_length || 128_000 and does not type or read top_provider.context_length. (src/agents/pi-embedded-runner/openrouter-model-capabilities.ts:177, 5be62e779b2e)
PR implementation uses endpoint context metadata: The PR diff bumps DISK_CACHE_VERSION to 3, adds top_provider.context_length, and resolves contextWindow from model.top_provider?.context_length ?? model.context_length ?? 128_000. (src/agents/pi-embedded-runner/openrouter-model-capabilities.ts:175, 76fcc362d24f)
Regression coverage matches the reported catalog shape: The PR adds tests for the reported nvidia/nemotron-3-super-120b-a12b:free model and verifies version-2 disk caches with stale context windows are not reused. (src/agents/pi-embedded-runner/openrouter-model-capabilities.test.ts:92, 76fcc362d24f)
Runtime consumes parsed dynamic capabilities: The OpenRouter provider plugin builds dynamic runtime models from getOpenRouterModelCapabilities, including capabilities.contextWindow and capabilities.maxTokens. (extensions/openrouter/index.ts:71, 5be62e779b2e)
Live OpenRouter catalog supports the root cause: A read-only live catalog lookup returned context_length: 1000000 and top_provider.context_length: 262144 for the reported model, matching the regression and PR proof.
Latest release has not shipped this fix: v2026.5.22 still contains disk cache version 2 and the old model.context_length || 128_000 parser expression, so the fix is not already in the latest release. (src/agents/pi-embedded-runner/openrouter-model-capabilities.ts:177, a374c3a5bfd5)

Likely related people:

steipete: Current main blame for the OpenRouter parser and dynamic model builder points to Peter Steinberger's recent carry-forward work, making this the clearest current-main routing signal. (role: recent area contributor; confidence: high; commits: b972ac194042, a374c3a5bfd5, 4a0f72866b4d; files: src/agents/pi-embedded-runner/openrouter-model-capabilities.ts, src/agents/pi-embedded-runner/openrouter-model-capabilities.test.ts, extensions/openrouter/index.ts)
Jinhao Dong: Git history shows the OpenRouter runtime capability detection feature was introduced in commit 8e4a1d87..., which is the feature this PR adjusts. (role: feature originator; confidence: medium; commits: 8e4a1d87e263; files: src/agents/pi-embedded-runner/openrouter-model-capabilities.ts, src/agents/pi-embedded-runner/openrouter-model-capabilities.test.ts)
Vincent Koc: The PR also adds declaration files for script helper modules, and recent history on those script helpers includes Vincent Koc's Windows npm shim hardening commit. (role: adjacent scripts owner; confidence: medium; commits: a56f4529720b; files: scripts/npm-runner.mjs, scripts/pnpm-runner.mjs, scripts/windows-cmd-helpers.mjs)

Codex review notes: model gpt-5.5, reasoning high; reviewed against 5be62e779b2e.

clawsweeper · 2026-05-24T12:47:06Z

ClawSweeper PR egg

✨ Hatched: 💎 rare Pearl Shellbean

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: 💎 rare.
Trait: purrs at green checks.
Image traits: location flaky test forest; accessory shell-shaped keyboard; palette moonlit blue and soft silver; mood sparkly; pose sitting proudly on a smooth stone; shell starlit enamel shell; lighting warm desk-lamp glow; background small review tokens.
Share on X: post this hatch
Copy: My PR egg hatched a 💎 rare Pearl Shellbean 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.

clawsweeper · 2026-05-24T12:47:20Z

🦞✅
ClawSweeper merged this PR after the passing review.

Source: clawsweeper[bot]
Feedback: structured ClawSweeper verdict: pass (sha=76fcc362d24f9d8d29b269f9d530cfdf4c71e4b2)
Merge status: merged by ClawSweeper automerge
Merged at: 2026-05-24T13:32:45Z
Merge commit: dd01a2e789fd

What merged:

The branch updates OpenRouter dynamic model capability parsing to prefer top_provider.context_length, bump ... sk cache version, adds regression coverage and a changelog entry, and adds script helper declaration files.
Reproducibility: yes. from source and live catalog evidence rather than an authenticated inference turn. Cur ... catalog currently reports a smaller endpoint-specific top_provider.context_length for the reported model.

Automerge notes:

PR branch already contained follow-up commit before automerge: fix(openrouter): use endpoint context limits
PR branch already contained follow-up commit before automerge: fix(clawsweeper): address review for automerge-openclaw-openclaw-8594…

The automerge loop is complete.

Automerge progress:

2026-05-24 12:56:21 UTC repair completed d193cae1f89a (no branch change) in 5m 56s Run: https://github.com/openclaw/clawsweeper/actions/runs/26361657158 command timed out after 300000ms: git fetch origin +refs/heads/clawsweeper/automerge-openclaw-openclaw-85949:refs/remotes/origin/clawsweeper/automerge-openclaw...

2026-05-24 12:56:20 UTC review queued d193cae1f89a (after no-op repair)

2026-05-24 12:56:22 UTC repair finished (blocked) in 5m 57s Run: https://github.com/openclaw/clawsweeper/actions/runs/26361657158 command timed out after 300000ms: git fetch origin +refs/heads/clawsweeper/automerge-openclaw-openclaw-85949:refs/remotes/origin/clawsweeper/automerge-openclaw...

2026-05-24 13:06:40 UTC repair queued d193cae1f89a (autonomous) Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845

2026-05-24 13:10:28 UTC repair started (running) in 1s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 automerge-openclaw-openclaw-85949

2026-05-24 13:10:44 UTC validation plan (passed) in 17s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 node scripts/run-vitest.mjs src/agents/pi-embedded-runner/openrouter-model-capabilities.test.ts src/agents/pi-embedded-runner/model.test.ts src/agents/openai-t...

2026-05-24 13:10:56 UTC Codex write preflight (passed) in 29s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 danger-full-access

2026-05-24 13:18:10 UTC Codex edit 1 23f4a82f56c8 (complete) in 7m 43s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 exit 0

2026-05-24 13:23:18 UTC validation and review 1 76fcc362d24f (base moved) in 12m 51s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 rebased

2026-05-24 13:22:05 UTC review queued d193cae1f89a (queued)

2026-05-24 13:23:22 UTC repair completed 76fcc362d24f (branch updated) in 12m 55s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 initial automerge rebase is delegated to Codex repair

2026-05-24 13:23:22 UTC review queued 76fcc362d24f (after repair)

2026-05-24 13:32:47 UTC automerge wait 76fcc362d24f (merged) in 22m 20s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 pull request already merged

2026-05-24 13:32:34 UTC review passed 76fcc362d24f (structured ClawSweeper verdict: pass (sha=76fcc362d24f9d8d29b269f9d530cfdf4c71e...)

2026-05-24 13:32:46 UTC merged 76fcc362d24f (merged by ClawSweeper automerge)

2026-05-24 13:32:48 UTC repair finished 76fcc362d24f (pushed) in 22m 21s Run: https://github.com/openclaw/clawsweeper/actions/runs/26362087845 repair_contributor_branch

…9 (1)

Summary: - The branch updates OpenRouter dynamic model capability parsing to prefer `top_provider.context_length`, bump ... sk cache version, adds regression coverage and a changelog entry, and adds script helper declaration files. - Reproducibility: yes. from source and live catalog evidence rather than an authenticated inference turn. Cur ... catalog currently reports a smaller endpoint-specific `top_provider.context_length` for the reported model. Automerge notes: - PR branch already contained follow-up commit before automerge: fix(openrouter): use endpoint context limits - PR branch already contained follow-up commit before automerge: fix(clawsweeper): address review for automerge-openclaw-openclaw-8594… Validation: - ClawSweeper review passed for head 76fcc36. - Required merge gates passed before the squash merge. Prepared head SHA: 76fcc36 Review: openclaw#86041 (comment) Co-authored-by: Andy Ye <35905412+TurboTheTurtle@users.noreply.github.com> Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: takhoffman Co-authored-by: takhoffman <781889+takhoffman@users.noreply.github.com>

* fix(compaction): preserve partial summary on mid-chain chunk failure (openclaw#82952) * fix(compaction): preserve partial summary on mid-chain chunk failure When summarizing multiple chunks, if a chunk fails after at least one chunk has already succeeded, return the partial summary instead of propagating the error and losing all summarization progress. Abort and timeout errors still propagate immediately. First-chunk failures still rethrow so the existing fallback path runs. Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca> * fix(compaction): use content array for assistant messages to match updated AgentMessage type * fix(compaction): use as-unknown-as-AgentMessage cast for assistant test fixtures --------- Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca> * fix(config): do not suppress recovery retry after failed backup restore (openclaw#85787) maybeRecoverSuspiciousConfigRead unconditionally recorded lastObservedSuspiciousSignature in health state even when restoredFromBackup was false (copyFile failed). The guard at resolveConfigReadRecoveryContext then prevented the same signature from ever being retried, permanently accepting the suspicious config on every subsequent launch. Only record the dedup signature when the backup restore actually succeeded. * fix(openrouter): use endpoint context limits (openclaw#86041) Summary: - The branch updates OpenRouter dynamic model capability parsing to prefer `top_provider.context_length`, bump ... sk cache version, adds regression coverage and a changelog entry, and adds script helper declaration files. - Reproducibility: yes. from source and live catalog evidence rather than an authenticated inference turn. Cur ... catalog currently reports a smaller endpoint-specific `top_provider.context_length` for the reported model. Automerge notes: - PR branch already contained follow-up commit before automerge: fix(openrouter): use endpoint context limits - PR branch already contained follow-up commit before automerge: fix(clawsweeper): address review for automerge-openclaw-openclaw-8594… Validation: - ClawSweeper review passed for head 76fcc36. - Required merge gates passed before the squash merge. Prepared head SHA: 76fcc36 Review: openclaw#86041 (comment) Co-authored-by: Andy Ye <35905412+TurboTheTurtle@users.noreply.github.com> Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: takhoffman Co-authored-by: takhoffman <781889+takhoffman@users.noreply.github.com> * fix(agents): log warnings instead of swallowing subagent errors (openclaw#82943) * fix: log subagent swallowed errors in hook emission and restore paths Wire createSubsystemLogger into the two silent catch blocks that discard errors during subagent lifecycle: 1. emitSubagentEndedHookOnce (subagent-registry-completion.ts): catch { return false } -> catch (err) { log.warn(...); return false } 2. restoreSubagentRunsOnce (subagent-registry.ts): catch { /* ignore */ } -> catch (err) { log.warn(...) } Both paths now log the error message before continuing, providing a diagnostic trail when hook emission or disk restore fails silently. Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca> * test(agents): keep provider test mocks current --------- Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca> Co-authored-by: Peter Steinberger <steipete@gmail.com> * fix(agents): match runtime policy entries when session provider is empty --------- Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca> Co-authored-by: Sebastien Tardif <SebTardif@ncf.ca> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Co-authored-by: Andy Ye <35905412+TurboTheTurtle@users.noreply.github.com> Co-authored-by: takhoffman <781889+takhoffman@users.noreply.github.com> Co-authored-by: Peter Steinberger <steipete@gmail.com> Co-authored-by: Brian Potter <brian@potterdigital.com>

Summary: - The branch updates OpenRouter dynamic model capability parsing to prefer `top_provider.context_length`, bump ... sk cache version, adds regression coverage and a changelog entry, and adds script helper declaration files. - Reproducibility: yes. from source and live catalog evidence rather than an authenticated inference turn. Cur ... catalog currently reports a smaller endpoint-specific `top_provider.context_length` for the reported model. Automerge notes: - PR branch already contained follow-up commit before automerge: fix(openrouter): use endpoint context limits - PR branch already contained follow-up commit before automerge: fix(clawsweeper): address review for automerge-openclaw-openclaw-8594… Validation: - ClawSweeper review passed for head 76fcc36. - Required merge gates passed before the squash merge. Prepared head SHA: 76fcc36 Review: openclaw#86041 (comment) Co-authored-by: Andy Ye <35905412+TurboTheTurtle@users.noreply.github.com> Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: takhoffman Co-authored-by: takhoffman <781889+takhoffman@users.noreply.github.com>

openclaw-barnacle Bot added size: S and removed size: XS labels May 24, 2026

clawsweeper Bot added rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. clawsweeper Tracked by ClawSweeper automation labels May 24, 2026

clawsweeper Bot mentioned this pull request May 24, 2026

fix(openrouter): use endpoint context limits #85949

Closed

openclaw-barnacle Bot removed the proof: supplied External PR includes structured after-fix real behavior proof. label May 24, 2026

clawsweeper Bot added status: 🚀 automerge armed This PR is in ClawSweeper's automerge lane. and removed status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. labels May 24, 2026

TurboTheTurtle and others added 4 commits May 24, 2026 13:23

fix(openrouter): use endpoint context limits

e90970e

fix(openrouter): use endpoint context limits

6e575c3

fix(clawsweeper): address review for automerge-openclaw-openclaw-8594…

3acc62c

…9 (1)

fix(openrouter): use endpoint context limits

76fcc36

clawsweeper Bot force-pushed the clawsweeper/automerge-openclaw-openclaw-85949 branch from d193cae to 76fcc36 Compare May 24, 2026 13:23

openclaw-barnacle Bot added the scripts Repository scripts label May 24, 2026

clawsweeper Bot merged commit dd01a2e into main May 24, 2026
106 of 107 checks passed

clawsweeper Bot deleted the clawsweeper/automerge-openclaw-openclaw-85949 branch May 24, 2026 13:32

github-actions Bot mentioned this pull request May 24, 2026

📡 Upstream Digest — 2026-05-24 14:56 UTC curtismercier/openclaw-mods#936

Open

frankhli843 mentioned this pull request May 25, 2026

maintenance: selective upstream openclaw sync 2026-05-25 gemmaclaw/gemmaclaw#235

Merged

2 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(openrouter): use endpoint context limits#86041

fix(openrouter): use endpoint context limits#86041
clawsweeper[bot] merged 4 commits into
mainfrom
clawsweeper/automerge-openclaw-openclaw-85949

clawsweeper Bot commented May 24, 2026

Uh oh!

clawsweeper Bot commented May 24, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 24, 2026

Uh oh!

clawsweeper Bot commented May 24, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

clawsweeper Bot commented May 24, 2026

Uh oh!

clawsweeper Bot commented May 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 24, 2026

Hatch command

Uh oh!

clawsweeper Bot commented May 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

clawsweeper Bot commented May 24, 2026 •

edited

Loading

clawsweeper Bot commented May 24, 2026 •

edited

Loading