fix(providers): use native reasoning mode for Gemini instead of tagged by 849261680 · Pull Request #89379 · openclaw/openclaw

849261680 · 2026-06-02T07:51:06Z

Summary

Gemini 2.5 Pro/Flash exposes reasoning via native thinkingParts through thinkingConfig.includeThoughts. The shared google-gemini family hook was selecting tagged reasoning output, which injected a <think>...</think>/<final>...</final> directive into the system prompt at the same time as the native thinking config was active.

The model then:

Emits a native thought:true part
Opens a <think> text block per the injected directive
Pivots to a tool call before closing </think>
Returns an empty post-tool assistant turn (content: [])

Gateway logs: incomplete turn detected: stopReason=stop payloads=0 - surfacing a spurious error to the user, while burning about 140K input tokens.

Changes

extensions/google/provider-registration.ts - override resolveReasoningOutputMode in the direct google provider to return "native", superseding the shared google-gemini family default of "tagged"
src/utils/provider-utils.ts - remove the now-dead BUILTIN_REASONING_OUTPUT_MODES entry keyed on "google-generative-ai": production callers pass the plugin id "google", not the transport API string, so the map entry was unreachable
docs/plugins/sdk-provider-plugins.md - document that direct Google uses native reasoning output while Gemini CLI keeps the tagged contract
Tests updated in both provider-utils.test.ts and extensions/google/index.test.ts

Real behavior proof

Behavior or issue addressed: Gemini 2.5 Pro/Flash with thinkingDefault: medium producing payloads=0 errors and spurious user-facing error messages on turns that include a tool call, as reported in Gemini text-tag reasoning conflicts with native thinking — produces unclosed <think>, empty post-tool turn, payloads=0 #69220.
Real environment tested: macOS arm64, Node v24.15.0, local source checkout on commit 2ed7ec9c3d; direct Google provider using auth profile google:default; model google/gemini-2.5-flash; embedded OpenClaw run with --thinking medium.
Exact steps or command run after this patch: pnpm openclaw agent --agent main --session-key pr-89379-gemini-proof --local --model google/gemini-2.5-flash --thinking medium --message 'Use an available local tool to run a harmless command that prints exactly OPENCLAW_GEMINI_PROOF, then reply with a one-sentence final answer that includes the command output. Do not search the web.' --json; node scripts/run-vitest.mjs src/utils/provider-utils.test.ts; node scripts/run-vitest.mjs extensions/google/index.test.ts; node ./node_modules/.bin/oxfmt --check docs/plugins/sdk-provider-plugins.md extensions/google/provider-registration.ts extensions/google/index.test.ts src/utils/provider-utils.ts src/utils/provider-utils.test.ts; pnpm docs:list; git diff --check.
Evidence after fix (screenshot, recording, terminal capture, console output, redacted runtime log, linked artifact, or copied live output): copied live output and redacted runtime log from the patched direct Gemini run:

{
  "payloads": [
    {
      "text": "The command output is OPENCLAW_GEMINI_PROOF.",
      "mediaUrl": null
    }
  ],
  "meta": {
    "provider": "google",
    "model": "gemini-2.5-flash",
    "requestShaping": { "authMode": "auth-profile", "thinking": "medium" },
    "toolSummary": { "calls": 1, "tools": ["exec"], "failures": 0 },
    "completion": { "stopReason": "stop", "finishReason": "stop" },
    "executionTrace": {
      "winnerProvider": "google",
      "winnerModel": "gemini-2.5-flash",
      "result": "success",
      "runner": "embedded",
      "fallbackUsed": false
    }
  },
  "sessionScan": {
    "events": [
      { "type": "model_change", "provider": "google", "modelId": "gemini-2.5-flash" },
      { "type": "assistant", "contentTypes": ["thinking", "toolCall"], "toolNames": ["exec"], "stopReason": "toolUse" },
      { "type": "toolResult", "ok": true },
      { "type": "assistant", "contentTypes": ["text"], "stopReason": "stop" }
    ],
    "hasThinkTag": false,
    "hasFinalTag": false
  }
}

Observed result after fix: the live Gemini turn used one exec tool call, returned a post-tool assistant final text payload, and did not hit payloads=0; the session scan found no injected <think> or <final> tags. Focused tests passed: provider-utils.test.ts 25/25, extensions/google/index.test.ts 11/11. Formatting/docs/diff checks passed.
What was not tested: Gemini 2.5 Pro live API path was tested with gemini-2.5-flash; Gemini 2.5 Pro specifically and Gemini 3 were not tested in this local credential profile. Older Gemini models without native thinking are expected unaffected because native mode only prevents OpenClaw from injecting text-tag reasoning directives.
Proof limitations or environment constraints: proof uses local auth-profile credentials and a local embedded OpenClaw run, not Crabbox, so no secret-bearing remote run was needed.

Tests and validation

pnpm openclaw agent --agent main --session-key pr-89379-gemini-proof --local --model google/gemini-2.5-flash --thinking medium --message 'Use an available local tool to run a harmless command that prints exactly OPENCLAW_GEMINI_PROOF, then reply with a one-sentence final answer that includes the command output. Do not search the web.' --json
node scripts/run-vitest.mjs src/utils/provider-utils.test.ts - 25/25 passed
node scripts/run-vitest.mjs extensions/google/index.test.ts - 11/11 passed
node ./node_modules/.bin/oxfmt --check docs/plugins/sdk-provider-plugins.md extensions/google/provider-registration.ts extensions/google/index.test.ts src/utils/provider-utils.ts src/utils/provider-utils.test.ts - clean
pnpm docs:list - completed
git diff --check - clean

Fixes #69220

clawsweeper · 2026-06-02T07:52:51Z

Codex review: needs maintainer review before merge. Reviewed June 2, 2026, 5:55 AM ET / 09:55 UTC.

Summary
The PR changes direct Google Gemini reasoning output to native mode, removes the core google-generative-ai tagged fallback, updates Google/CLI regression tests, and documents the direct-vs-CLI reasoning contract.

PR surface: Source -10, Tests +16, Docs +7. Total +13 across 5 files.

Reproducibility: yes. from source and the linked live proof: current main can combine direct Gemini thinkingConfig.includeThoughts with the tagged <think>/<final> prompt directive. I did not run a live current-main repro in this read-only review.

Review metrics: 1 noteworthy metric.

Reasoning output modes: 1 direct provider changed to native; 1 CLI provider preserved as tagged. The quantified behavior split is the compatibility-sensitive part of the PR: direct Google changes output contract while Gemini CLI keeps the old tagged contract.

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:

[P2] Optional: add Gemini 2.5 Pro or Gemini 3 live tool-call proof if maintainers want coverage beyond Flash before merge.

Risk before merge

[P1] Merging intentionally changes shipped direct Google Gemini output behavior from tagged <think>/<final> prompting to native thought parts, so maintainers should accept that older direct-Google final-tag enforcement is no longer the default.
[P1] The supplied live proof covers google/gemini-2.5-flash; Gemini 2.5 Pro and Gemini 3 are source-supported by the same path but not separately live-tested in the PR body.

Maintainer options:

Accept the native-mode tradeoff (recommended)
Merge with the current provider-local override, CLI regression test, docs update, and live Gemini 2.5 Flash proof if maintainers accept the direct-Google output-mode change.
Broaden live provider proof first
Ask for one additional Gemini 2.5 Pro or Gemini 3 tool-call run with thinking enabled if maintainers want live coverage beyond the Flash path before merge.

Next step before merge

No ClawSweeper repair lane is needed because review found no concrete patch defect; maintainers can handle merge/risk acceptance through the normal PR flow.

Security
Cleared: No concrete security or supply-chain concern found; the diff changes provider output-mode logic, tests, and docs without adding dependencies, secrets handling, CI, or code-download behavior.

Review details

Best possible solution:

Land the direct-Google native override with the CLI tagged invariant and docs/tests intact, after maintainers accept the provider-output compatibility tradeoff.

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

Yes, from source and the linked live proof: current main can combine direct Gemini thinkingConfig.includeThoughts with the tagged <think>/<final> prompt directive. I did not run a live current-main repro in this read-only review.

Is this the best way to solve the issue?

Yes; the provider-local override is the narrowest maintainable fix because direct Google consumes native thought parts while Gemini CLI still needs the shared tagged text-output contract. Changing the shared GOOGLE_GEMINI_PROVIDER_HOOKS would be too broad.

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against e24582d53cbf.

Label changes

Label changes:

add proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes copied live output and a redacted runtime/session scan from a patched Gemini 2.5 Flash --thinking medium tool-call run showing a final payload, no payloads=0, and no injected tags.
add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🦞 diamond lobster and patch quality is 🐚 platinum hermit.
add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes copied live output and a redacted runtime/session scan from a patched Gemini 2.5 Flash --thinking medium tool-call run showing a final payload, no payloads=0, and no injected tags.
remove rating: 🧂 unranked krab: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.
remove status: 📣 needs proof: Current PR status label is status: 👀 ready for maintainer look.

Label justifications:

P1: The linked bug affects real direct-Gemini agent/tool-call replies by producing empty post-tool payloads and user-facing errors.
merge-risk: 🚨 compatibility: The PR intentionally changes the shipped direct Google Gemini reasoning-output contract from tagged prompt directives to native thought parts.
merge-risk: 🚨 message-delivery: The changed output mode sits on the path that determines whether post-tool assistant text becomes a visible payload or an empty reply.
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 copied live output and a redacted runtime/session scan from a patched Gemini 2.5 Flash --thinking medium tool-call run showing a final payload, no payloads=0, and no injected tags.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes copied live output and a redacted runtime/session scan from a patched Gemini 2.5 Flash --thinking medium tool-call run showing a final payload, no payloads=0, and no injected tags.

Evidence reviewed

PR surface:

Source -10, Tests +16, Docs +7. Total +13 across 5 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	2	6	16	-10
Tests	2	20	4	+16
Docs	1	8	1	+7
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	5	34	21	+13

What I checked:

Current main still injects tagged reasoning for Google-family providers: On current main, buildProviderReplayFamilyHooks({ family: "google-gemini" }) returns resolveReasoningOutputMode as tagged, and the direct Google provider spreads GOOGLE_GEMINI_PROVIDER_HOOKS without overriding it. (src/plugin-sdk/provider-model-shared.ts:238, e24582d53cbf)
Current main maps the transport API string to tagged as a fallback: BUILTIN_REASONING_OUTPUT_MODES still contains "google-generative-ai": "tagged" on current main, so the PR's deletion is not already implemented. (src/utils/provider-utils.ts:10, e24582d53cbf)
Runtime callers pass provider id and carry model API separately: The embedded runner calls isReasoningTagProvider(params.provider, ...) and passes modelApi: params.model.api only as context, supporting the PR's claim that provider ids own the output-mode decision. (src/agents/embedded-agent-runner/run/attempt.ts:1698, e24582d53cbf)
Direct Google already requests native thinking parts: The direct Google transport builds generationConfig.thinkingConfig with includeThoughts: true for thinking-enabled Gemini requests, so keeping tagged prompt directives at the same time creates overlapping reasoning contracts. (extensions/google/transport-stream.ts:441, e24582d53cbf)
Direct Google consumes native thought parts: The direct Google stream parser treats part.thought === true as a thinking block and persists thoughtSignature, matching native Gemini output rather than text-tag output. (extensions/google/transport-stream.ts:1251, e24582d53cbf)
Dependency contract supports native thought fields: The published @google/[email protected] types expose GenerateContentConfig.thinkingConfig, ThinkingConfig.includeThoughts, and Part.thought / Part.thoughtSignature, matching the PR's native-mode rationale. (extensions/google/package.json:8, e24582d53cbf)

Likely related people:

steipete: Recent GitHub commit history shows multiple Google provider and Gemini transport changes, including moving Gemini transport into the plugin and recovering Gemini tool-call thought signatures. (role: recent Google provider/runtime contributor; confidence: high; commits: 85826c83e4a1, 8dd91b14d3c0, 9ead0ae9219e; files: extensions/google/provider-registration.ts, extensions/google/transport-stream.ts)
vincentkoc: Commit 356110c52f79ff13e698cbae518a5c2ff8795b5d added GOOGLE_GEMINI_PROVIDER_HOOKS and shared the Gemini hook bundle between direct Google and Gemini CLI. (role: shared Google Gemini hook introducer; confidence: high; commits: 356110c52f79; files: extensions/google/provider-hooks.ts, extensions/google/provider-registration.ts, extensions/google/gemini-cli-provider.ts)
shakkernerd: Commit 590655472b4c2e20d273c746826b8c3fa904ff46 introduced the built-in reasoning provider fast path and the original google-generative-ai tagged fallback that this PR removes. (role: provider-utils reasoning fallback introducer; confidence: high; commits: 590655472b4c; files: src/utils/provider-utils.ts, src/utils/provider-utils.test.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.

… CLI tagged Gemini 2.5+ delivers reasoning via native thinkingParts (thinkingConfig. includeThoughts). Having tagged mode active at the same time injects a <think>…</think>/<final>…</final> directive into the system prompt; the model opens a <think> block before a tool call, never closes it, and returns an empty post-tool turn (content:[], payloads=0 error, openclaw#69220). Fix: override resolveReasoningOutputMode in buildGoogleProvider() only — not in the shared GOOGLE_GEMINI_PROVIDER_HOOKS. The Gemini CLI backend (google-gemini-cli) runs gemini --output-format json and parses a text response field, not native thought parts; it must stay on tagged mode. A regression test confirms google-gemini-cli remains "tagged". Also remove the dead BUILTIN_REASONING_OUTPUT_MODES entry keyed on "google-generative-ai" from provider-utils.ts — that string is only ever the transport model.api value, never the provider id passed to resolveReasoningOutputMode, so the map was unreachable. Fixes openclaw#69220

849261680 · 2026-06-02T08:07:27Z

Addressed ClawSweeper's P1 finding: the resolveReasoningOutputMode: () => "native" override has been moved from the shared GOOGLE_GEMINI_PROVIDER_HOOKS to buildGoogleProvider() only.

google-gemini-cli continues to inherit "tagged" from the google-gemini family hook (no behavior change for CLI). A regression test (keeps google-gemini-cli on tagged reasoning mode) is now added to index.test.ts to lock this invariant.

@clawsweeper re-review

clawsweeper · 2026-06-02T08:07:31Z

🦞🧹
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/26807006352
Updated: 2026-06-02T08:15:13.875Z

849261680 · 2026-06-02T09:44:09Z

Updated the PR with the direct Google/Gemini docs alignment and a redacted live Gemini 2.5 Flash --thinking medium tool-call proof from the patched branch. The live run shows one exec tool call, a normal post-tool final payload, no payloads=0, and no injected <think> / <final> tags in the session scan.

@clawsweeper re-review

clawsweeper · 2026-06-02T09:44:12Z

🦞🧹
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.

849261680 · 2026-06-02T09:46:43Z

The Real behavior proof section is now a proper ## Real behavior proof heading and the repo policy script evaluates the current PR body as passed. GitHub has also applied proof: supplied after the updated body.

@clawsweeper re-review

clawsweeper · 2026-06-02T09:46:46Z

🦞👀
ClawSweeper picked this up.

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

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/26811932407
Updated: 2026-06-02T09:55:49.780Z

steipete · 2026-06-02T11:45:50Z

Maintainer verification for exact head 459dad98e2744be9ac2a81fbd829c8a57579b658:

Reviewed and kept the contributor's live direct-Gemini proof: google/gemini-2.5-flash with --thinking medium used an exec tool, returned final text, and did not produce the payloads=0 incomplete-turn failure from Gemini text-tag reasoning conflicts with native thinking — produces unclosed <think>, empty post-tool turn, payloads=0 #69220.
Added maintainer fixup e56ff80ffb so google-antigravity text-style hook aliases remain tagged while native Gemini transports stay native.
Added maintainer fixup 22025556a1 so provider-only direct google / google-vertex reasoning checks default to native when modelApi is not yet threaded, keeping final-tag filtering aligned with prompt mode.
Added maintainer fixup 459dad98e2 to import ProviderReasoningOutputModeContext from the plugin-entry SDK barrel, matching the exported type surface and fixing the type lanes.
Ran node scripts/run-vitest.mjs extensions/google/index.test.ts src/utils/provider-utils.test.ts: provider-utils.test.ts 25/25 and extensions/google/index.test.ts 11/11 passed.
Ran node ./node_modules/.bin/oxfmt --check docs/plugins/sdk-provider-plugins.md extensions/google/provider-registration.ts extensions/google/index.test.ts src/utils/provider-utils.ts src/utils/provider-utils.test.ts.
Ran git diff --check.
Ran pnpm docs:list.
Ran pnpm tsgo:extensions and pnpm check:test-types after the maintainer import fix.
Ran branch autoreview against origin/main on the final branch shape; no accepted/actionable findings remained.
Exact-head GitHub checks are green: 141 pass, 25 skipped.

Known proof gap: the live Gemini run was from the contributor proof before the maintainer alias/type fixups; those fixups were covered with source assertions, focused tests, type lanes, autoreview, and exact-head CI.

openclaw#89379) * fix(providers): use native reasoning mode for direct Gemini API, keep CLI tagged Gemini 2.5+ delivers reasoning via native thinkingParts (thinkingConfig. includeThoughts). Having tagged mode active at the same time injects a <think>…</think>/<final>…</final> directive into the system prompt; the model opens a <think> block before a tool call, never closes it, and returns an empty post-tool turn (content:[], payloads=0 error, openclaw#69220). Fix: override resolveReasoningOutputMode in buildGoogleProvider() only — not in the shared GOOGLE_GEMINI_PROVIDER_HOOKS. The Gemini CLI backend (google-gemini-cli) runs gemini --output-format json and parses a text response field, not native thought parts; it must stay on tagged mode. A regression test confirms google-gemini-cli remains "tagged". Also remove the dead BUILTIN_REASONING_OUTPUT_MODES entry keyed on "google-generative-ai" from provider-utils.ts — that string is only ever the transport model.api value, never the provider id passed to resolveReasoningOutputMode, so the map was unreachable. Fixes openclaw#69220 * docs: clarify Gemini reasoning output modes * fix(google): keep Antigravity reasoning tagged * fix(google): default direct reasoning checks to native * fix(google): import reasoning context from plugin entry --------- Co-authored-by: Peter Steinberger <[email protected]>

openclaw-barnacle Bot added extensions: google size: XS triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels Jun 2, 2026

849261680 force-pushed the fix/gemini-reasoning-output-mode branch from 58b1e75 to 76b1356 Compare June 2, 2026 08:07

docs: clarify Gemini reasoning output modes

2ed7ec9

openclaw-barnacle Bot added the docs Improvements or additions to documentation label Jun 2, 2026

openclaw-barnacle Bot added proof: supplied External PR includes structured after-fix real behavior proof. and removed triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels Jun 2, 2026

steipete added 2 commits June 2, 2026 04:32

fix(google): keep Antigravity reasoning tagged

e56ff80

fix(google): default direct reasoning checks to native

2202555

openclaw-barnacle Bot added size: S and removed size: XS labels Jun 2, 2026

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label Jun 2, 2026

fix(google): import reasoning context from plugin entry

459dad9

steipete merged commit b1bdc29 into openclaw:main Jun 2, 2026
166 checks passed

github-actions Bot mentioned this pull request Jun 2, 2026

📡 Upstream Digest — 2026-06-02 13:53 UTC curtismercier/openclaw-mods#1002

Open

This was referenced Jun 14, 2026

Gemini 2.5 Pro: textSignature bloat + think tags + mixed text/tool causes session failures #48709

Open

WhatsApp: Gemini thinking tags (<think>/<final>) still leak after #6328 fix #87712

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(providers): use native reasoning mode for Gemini instead of tagged#89379

fix(providers): use native reasoning mode for Gemini instead of tagged#89379
steipete merged 5 commits into
openclaw:mainfrom
849261680:fix/gemini-reasoning-output-mode

849261680 commented Jun 2, 2026 •

edited by steipete

Loading

Uh oh!

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

Uh oh!

849261680 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

Uh oh!

849261680 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

Uh oh!

849261680 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

Uh oh!

steipete commented Jun 2, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

849261680 commented Jun 2, 2026 • edited by steipete Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Real behavior proof

Tests and validation

Uh oh!

clawsweeper Bot commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

849261680 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

849261680 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

849261680 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

steipete commented Jun 2, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

849261680 commented Jun 2, 2026 •

edited by steipete

Loading

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading