fix(status): ignore stale context after model switch by hxy91819 · Pull Request #93306 · openclaw/openclaw

hxy91819 · 2026-06-15T13:46:11Z

Purpose

Manual session model switches should make /status report the context window for the newly selected model. The current status formatter can still trust a transcript/runtime snapshot from the previous model, so a session pinned to ollama-cloud/glm-5.1 can continue to display the old ollama-cloud/deepseek-v4-pro 1.0m context window.

Summary

Keep /status context-window display aligned with the session-selected model after manual model switches.
Trust explicit runtime context tokens only when the runtime snapshot has fallback or equivalent-runtime provenance.
Add regression coverage for both the stale-runtime case and config-backed runtime aliases that should still keep their runtime window.

Root Cause

src/status/status-message.ts treated runtimeContextTokens as authoritative whenever the runtime model differed from the selected model. That is correct for real fallback or equivalent runtime alias paths, but wrong after a manual session model switch because the runtime snapshot can describe the previous transcript model.

Real Behavior Proof

Behavior addressed: /status now uses the session-selected model's configured context window after a manual model switch instead of pinning the display to stale runtime context tokens from the previous model.

Real environment tested: Crabbox Docker provider, lease cbx_9a713b50341d, clean synced PR branch checkout in node:22-bookworm.

Exact steps or command run after this patch: node scripts/crabbox-wrapper.mjs run --provider docker --docker-image node:22-bookworm --idle-timeout 90m --ttl 240m --timing-json --shell -- "corepack enable && corepack pnpm install --frozen-lockfile && node scripts/run-vitest.mjs src/status/status-message.test.ts"

Evidence after fix: src/status/status-message.test.ts builds the stale manual-switch status input where the selected model is ollama-cloud/glm-5.1, the previous runtime snapshot is ollama-cloud/deepseek-v4-pro, and the stale runtime context is 1_000_000.

Observed result after fix: Crabbox passed src/status/status-message.test.ts with 1 test file and 4 tests. The regression test asserts Session selected: ollama-cloud/glm-5.1, Context: 128k/200k, and not Context: 128k/1.0m.

What was not tested: No live Telegram/openclaw2 package upgrade was rerun from this PR branch; the focused proof covers the status-message decision path that formats the incorrect context window.

Verification

pnpm check:test-types
node scripts/run-vitest.mjs src/status/status-message.test.ts
.agents/skills/autoreview/scripts/autoreview --mode local
git diff --check
Crabbox Docker proof: provider docker, lease cbx_9a713b50341d, exit 0.

chatgpt-codex-connector · 2026-06-15T13:46:23Z

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

clawsweeper · 2026-06-15T13:48:48Z

Codex review: passed. Reviewed June 15, 2026, 3:32 PM ET / 19:32 UTC.

Summary
The PR changes /status context-window selection to ignore stale runtime snapshots after manual model switches while preserving fallback/runtime-alias context windows.

PR surface: Source +6, Tests +128. Total +134 across 2 files.

Reproducibility: yes. source-reproducible: current main trusts explicit runtime context before checking fallback provenance when active and selected models differ. I did not run a local failing repro, but the PR fixture models the stale prior-runtime state directly.

Review metrics: none identified.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🌊 off-meta tidepool
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:

none.

Next step before merge

[P2] No repair lane is needed; the PR has no actionable review finding and can continue through normal maintainer/automerge gates.

Security
Cleared: The diff only changes TypeScript status-formatting logic and colocated tests; it does not touch dependencies, workflows, packaging, secrets, or code-execution surfaces.

Review details

Best possible solution:

Land the focused status formatter guard with its regression coverage once the normal maintainer and automerge gates accept the exact head SHA.

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

Yes, source-reproducible: current main trusts explicit runtime context before checking fallback provenance when active and selected models differ. I did not run a local failing repro, but the PR fixture models the stale prior-runtime state directly.

Is this the best way to solve the issue?

Yes. Moving explicit runtime-token trust behind the fallback/equivalent-runtime provenance check is the narrow formatter-side fix, and passing args.config preserves configured runtime aliases.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against b3128ba93def.

Label changes

Label changes:

add status: 🚀 automerge armed: This PR is in ClawSweeper's automerge lane. Not applicable: The external-contributor proof gate is not applicable because the PR is maintainer-labeled; the body still records focused Crabbox Docker proof for the status-message test path.
remove status: 👀 ready for maintainer look: Current PR status label is status: 🚀 automerge armed.

Label justifications:

P2: This fixes a bounded user-facing /status display bug after manual model switches with limited blast radius.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
status: 🚀 automerge armed: This PR is in ClawSweeper's automerge lane. Not applicable: The external-contributor proof gate is not applicable because the PR is maintainer-labeled; the body still records focused Crabbox Docker proof for the status-message test path.

Evidence reviewed

PR surface:

Source +6, Tests +128. Total +134 across 2 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	1	19	13	+6
Tests	1	129	1	+128
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	2	148	14	+134

What I checked:

Current-main stale path: On current main, buildStatusMessage enters the runtime-diff branch and returns explicitRuntimeContextTokens before the fallback/equivalence provenance check can reject stale runtime metadata. (src/status/status-message.ts:799, b3128ba93def)
PR-head fix: At PR head, runtimeSnapshotHasFallbackProvenance is computed first with args.config, and explicit runtime context tokens are trusted only after fallback or equivalent-runtime provenance is established. (src/status/status-message.ts:788, f14fda427943)
Regression coverage: The PR adds tests for the stale manual-switch case and for config-backed runtime aliases that should retain the runtime context window. (src/status/status-message.test.ts:35, f14fda427943)
Caller contract: buildStatusText already filters runtime context tokens with the same fallback/equivalent-runtime provenance before delegating to buildStatusMessage, so the PR aligns the lower-level formatter with its caller. (src/status/status-text.ts:557, b3128ba93def)
Latest release still lacks the fix: The latest release tag still has the older status-message context branch and does not contain the PR head commits, so the requested behavior is not already shipped. (src/status/status-message.ts:697, 8c802aa68351)
Related merged provenance: The current status context-window branch comes from the recently merged stale-context-window repair, and this PR is a focused follow-up to that same surface. (src/status/status-message.ts:799, 4029fbd2b238)

Likely related people:

hxy91819: Authored the merged stale status context-window repair in fix(status): avoid stale session context windows #93220 and this focused follow-up on the same formatter path. (role: recent area contributor; confidence: high; commits: 4029fbd2b238, 570c8d1828c5, f14fda427943; files: src/status/status-message.ts, src/status/status-text.ts, src/status/status-message.test.ts)
Josh Lehman: Recent local history shows adjacent status module work in status-message.ts and status-text.ts, including plugin-health and transcript-reader changes. (role: recent adjacent status contributor; confidence: medium; commits: 10a4c7c10b6a, 8ded75628437; files: src/status/status-message.ts, src/status/status-text.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.

hxy91819 · 2026-06-15T14:21:18Z

@clawsweeper automerge

clawsweeper · 2026-06-15T14:21:22Z

🦞👀
ClawSweeper picked this up.

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

hxy91819 · 2026-06-16T00:39:52Z

@clawsweeper automerge

clawsweeper · 2026-06-16T00:39:55Z

🦞✅
ClawSweeper merged this PR after the passing review.

Source: clawsweeper[bot]
Feedback: structured ClawSweeper verdict: pass (sha=f14fda4279435f1407ee146bba97e03e2283255a)
Merge status: merged by ClawSweeper automerge
Merged at: 2026-06-16T15:05:52Z
Merge commit: f046d7aa23df

What merged:

The PR changes /status context-window selection to ignore stale runtime snapshots after manual model switches while preserving fallback/runtime-alias context windows.
PR surface: Source +6, Tests +128. Total +134 across 2 files.
Reproducibility: yes. source-reproducible: current main trusts explicit runtime context before checking fall ... fer. I did not run a local failing repro, but the PR fixture models the stale prior-runtime state directly.

Automerge notes:

PR branch already contained follow-up commit before automerge: test(status): make context fixtures type-correct

The automerge loop is complete.

Automerge progress:

2026-06-16 00:40:51 UTC review queued f14fda427943 (queued)

2026-06-16 14:10:44 UTC review passed f14fda427943 (structured ClawSweeper verdict: pass (sha=f14fda4279435f1407ee146bba97e03e22832...)

2026-06-16 15:05:56 UTC merged f14fda427943 (merged by ClawSweeper automerge)

clawsweeper · 2026-06-16T13:14:32Z

🦞✅
ClawSweeper is pausing this repair loop for human review.

Source: clawsweeper[bot]
Reason: - [P2] No repair lane is needed; the PR has no actionable review finding and can continue through normal maintainer/automerge gates.; Cleared: The diff only changes TypeScript status-formatting logic and colocated tests; it does not touch dependencies, workflows, packaging, secrets, or code-execution surfaces. (sha=f14fda4279435f1407ee146bba97e03e2283255a)

Why human review is needed:
This item has security-sensitive risk. ClawSweeper is pausing instead of making an autonomous change that could affect trust, credentials, permissions, or exposure.

What the maintainer can do as a next step:
If the maintainer accepts the current risk and wants ClawSweeper to continue merge gates, comment @clawsweeper approve. If the security-sensitive detail still needs changes, describe the safe path or push the fix, then comment @clawsweeper automerge. If the risk should not be automated, keep the PR paused for manual review or comment @clawsweeper stop.

I added clawsweeper:human-review and left the final call with a maintainer.

Summary: - The PR changes `/status` context-window selection to ignore stale runtime snapshots after manual model switches while preserving fallback/runtime-alias context windows. - PR surface: Source +6, Tests +128. Total +134 across 2 files. - Reproducibility: yes. source-reproducible: current main trusts explicit runtime context before checking fall ... fer. I did not run a local failing repro, but the PR fixture models the stale prior-runtime state directly. Automerge notes: - PR branch already contained follow-up commit before automerge: test(status): make context fixtures type-correct Validation: - ClawSweeper review passed for head f14fda4. - Required merge gates passed before the squash merge. Prepared head SHA: f14fda4 Review: openclaw#93306 (comment) Co-authored-by: Mason Huang <[email protected]> Approved-by: hxy91819

fix(status): ignore stale context after model switch

570c8d1

openclaw-barnacle Bot added size: S maintainer Maintainer-authored PR labels Jun 15, 2026

test(status): make context fixtures type-correct

f14fda4

clawsweeper Bot added clawsweeper:automerge Maintainer opted this PR into bounded ClawSweeper-reviewed automerge clawsweeper:human-review Needs maintainer review before ClawSweeper can continue labels Jun 16, 2026

clawsweeper Bot merged commit f046d7a into main Jun 16, 2026
217 of 223 checks passed

clawsweeper Bot deleted the codex/status-context-window-model-switch branch June 16, 2026 15:05

clawsweeper Bot mentioned this pull request Jun 18, 2026

fix: refresh status context window after model switch #80957

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(status): ignore stale context after model switch#93306

fix(status): ignore stale context after model switch#93306
clawsweeper[bot] merged 2 commits into
mainfrom
codex/status-context-window-model-switch

hxy91819 commented Jun 15, 2026 •

edited

Loading

Uh oh!

chatgpt-codex-connector Bot commented Jun 15, 2026

Uh oh!

clawsweeper Bot commented Jun 15, 2026 •

edited

Loading

Uh oh!

hxy91819 commented Jun 15, 2026

Uh oh!

clawsweeper Bot commented Jun 15, 2026

Uh oh!

hxy91819 commented Jun 16, 2026

Uh oh!

clawsweeper Bot commented Jun 16, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented Jun 16, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Uh oh!

Conversation

hxy91819 commented Jun 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Purpose

Summary

Root Cause

Real Behavior Proof

Verification

Uh oh!

chatgpt-codex-connector Bot commented Jun 15, 2026

Uh oh!

clawsweeper Bot commented Jun 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

hxy91819 commented Jun 15, 2026

Uh oh!

clawsweeper Bot commented Jun 15, 2026

Uh oh!

hxy91819 commented Jun 16, 2026

Uh oh!

clawsweeper Bot commented Jun 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented Jun 16, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

hxy91819 commented Jun 15, 2026 •

edited

Loading

clawsweeper Bot commented Jun 15, 2026 •

edited

Loading

clawsweeper Bot commented Jun 16, 2026 •

edited

Loading