[AI] fix(memory): prevent empty-string expectedModel in resolveMemory… by xydt-tanshanshan · Pull Request #91691 · openclaw/openclaw

xydt-tanshanshan · 2026-06-09T12:18:12Z

Summary

resolveMemoryIndexIdentityState computes expectedModel from provider.model, but when provider.model is an empty string (as happens when resolveEmbeddingProviderFallbackModel returns "" for an unregistered adapter, or when createWithAdapter returned empty model before PR #90042), the expected model becomes "" — causing a permanent identity mismatch that blocks vector search indefinitely.

This PR fixes two code paths that could produce empty-string expectedModel:

Identity consumer layer (manager-reindex-state.ts:123): Replace params.provider ? params.provider.model : "fts-only" with params.provider?.model?.trim() || "fts-only". Empty or whitespace-only provider.model now falls back to "fts-only" — the same safe default used when no provider is configured.
Identity computation layer (manager-sync-ops.ts:420): Change resolveEmbeddingProviderFallbackModel(provider, "", cfg) to resolveEmbeddingProviderFallbackModel(provider, "fts-only", cfg). When the adapter IS registered (e.g., "openai" → defaultModel "text-embedding-3-small"), the resolved default model is used. When the adapter is NOT registered (e.g., "google" before the Google plugin loads), fallbackSourceModel = "fts-only" is returned instead of "".

This two-layer approach addresses the ClawSweeper P1 finding: when a configured provider has an empty model, the resolved adapter default model (e.g., "text-embedding-3-small") is used for identity comparison. "fts-only" only appears when both the adapter and the configured model are unavailable — matching the no-provider path semantics.

Change Type

Bug fix

Scope

Memory / storage

Linked Issue

Related to [Bug]: memorySearch provider silently resets to "openai" after upgrade to 2026.6.1, causing permanent Dirty index and vector search outage #90787 (prevents empty-string expectedModel in identity check; provider auto→openai migration decision is a separate product issue)
Related to Gateway memory_search index stuck dirty: provider.model empty during boot overwrites correct index identity #90042 (provider.model backfill in createWithAdapter; this PR protects the identity check even when backfill hasn't happened yet or the adapter is unavailable)

Motivation

Users upgrading from 2026.5.26 to 2026.6.1 who previously used Google/Gemini as their memorySearch provider see Dirty: yes permanently with reason: "index was built for model gemini-embedding-001, expected " — the empty expected field makes the index unrecoverable. PR #90042 fixed the root cause at createWithAdapter (provider initialization), but the identity check can still produce empty-string expectedModel before provider init completes, or when the adapter registry hasn't loaded yet.

Changes

File: extensions/memory-core/src/memory/manager-reindex-state.ts

Replace params.provider ? params.provider.model : "fts-only" with params.provider?.model?.trim() || "fts-only" — guard against empty/whitespace provider.model at the identity comparison consumer

File: extensions/memory-core/src/memory/manager-sync-ops.ts

Change resolveEmbeddingProviderFallbackModel(this.settings.provider, "", this.cfg) to resolveEmbeddingProviderFallbackModel(this.settings.provider, "fts-only", this.cfg) — use "fts-only" as fallbackSourceModel instead of "" so unregistered adapters produce a readable fallback instead of empty string

File: extensions/memory-core/src/memory/manager-reindex-state.test.ts

Add 3 tests covering empty-string, whitespace-only, and non-fts mismatch scenarios

No new imports, no config changes, no default surface changes.

Verification

extensions/memory-core/src/memory/manager-reindex-state.test.ts — 12 passed (9 original + 3 new)
extensions/memory-core/src/memory/embeddings.test.ts — 8 passed (original suite, PR Gateway memory_search index stuck dirty: provider.model empty during boot overwrites correct index identity #90042's 2 backfill tests are on a separate branch)

Real Behavior Proof

Behavior addressed: expectedModel in resolveMemoryIndexIdentityState can become an empty string "" when provider.model is empty, causing permanent identity mismatch ("index was built for model X, expected ") and blocking vector search recovery.

Real environment tested: Linux (Ubuntu), Node.js 22.19+, local checkout on branch fix/memory-identity-empty-model-90787.

Exact steps or command run after this patch:

node scripts/run-vitest.mjs extensions/memory-core/src/memory/manager-reindex-state.test.ts --run
node /tmp/identity-proof.mjs
node /tmp/fallback-proof.mjs
http_proxy=http://proxyhk.zte.com.cn:80 https_proxy=http://proxyhk.zte.com.cn:80 pnpm openclaw memory status --deep

Evidence after fix: Direct runtime proof demonstrating the empty-string provider.model path changed after the fix. Two focused scripts exercise the exact code paths that this PR changes:

resolveMemoryIndexIdentityState expectedModel comparison (node /tmp/identity-proof.mjs): proves the identity consumer layer now produces "fts-only" instead of empty string for empty provider.model, while preserving behavior for normal values:

=== resolveMemoryIndexIdentityState: expectedModel BEFORE vs AFTER fix ===

Case: provider.model = "" (issue #90787 exact scenario)
  BEFORE: expectedModel = ""
  AFTER:  expectedModel = "fts-only"
  Changed: YES >>> empty string eliminated

Case: provider.model = "  " (whitespace)
  BEFORE: expectedModel = "  "
  AFTER:  expectedModel = "fts-only"
  Changed: YES >>> empty string eliminated

Case: provider.model = "text-embedding-3-small" (normal)
  BEFORE: expectedModel = "text-embedding-3-small"
  AFTER:  expectedModel = "text-embedding-3-small"
  Changed: NO >>> behavior preserved

Case: provider = null (no-provider path)
  BEFORE: expectedModel = "fts-only"
  AFTER:  expectedModel = "fts-only"
  Changed: NO >>> behavior preserved

Case: provider.model = "gemini-embedding-001" (google)
  BEFORE: expectedModel = "gemini-embedding-001"
  AFTER:  expectedModel = "gemini-embedding-001"
  Changed: NO >>> behavior preserved

resolveEmbeddingProviderFallbackModel fallbackSourceModel comparison (node /tmp/fallback-proof.mjs): proves the pre-init configured-provider fallback now uses "fts-only" for unregistered adapters instead of empty string, while preserving resolved defaultModel for registered adapters:

=== resolveEmbeddingProviderFallbackModel: fallbackSourceModel BEFORE vs AFTER ===

Case: openai adapter registered, empty fallbackSourceModel (BEFORE)
  Result: "text-embedding-3-small"

Case: openai adapter registered, fts-only fallbackSourceModel (AFTER)
  Result: "text-embedding-3-small"

Case: google adapter NOT registered, empty fallbackSourceModel (BEFORE)
  Result: ""

Case: google adapter NOT registered, fts-only fallbackSourceModel (AFTER)
  Result: "fts-only"

Unit test (terminal output from node scripts/run-vitest.mjs): directly tests resolveMemoryIndexIdentityState with provider.model = "" → expectedModel = "fts-only" (not empty string):

 ✓ falls back to fts-only when provider.model is an empty string
 ✓ reports mismatch when empty-string expected model is compared to a non-fts index
 ✓ falls back to fts-only when provider.model is whitespace-only

Before-fix comparison (from issue #90787 reporter, Rocky Linux, v2026.6.1):

Field	Before (issue #90787)	After (this patch)
`expectedModel`	`""` (empty string)	adapter defaultModel or `"fts-only"` (safe fallback)
Identity reason	`"index was built for model gemini-embedding-001, expected "` (trailing empty string)	`"index was built for model gemini-embedding-001, expected text-embedding-3-small"` (adapter default) or `"expected fts-only"` (unregistered adapter)
Recovery path	None (permanent Dirty, no rebuild clears it)	Reindex with correct provider → Dirty: no

Observed result after fix: Two-layer protection ensures expectedModel is never an empty string. When the adapter IS registered (e.g., "openai" → "text-embedding-3-small"), the resolved default model is used for identity comparison. When the adapter is NOT registered, "fts-only" is used — matching the no-provider path semantics and producing a readable mismatch reason that points the user toward a reindex instead of a broken permanent-stuck state.

What was not tested: The provider auto→openai migration decision (separate product issue with needs-product-decision label). This PR only fixes the identity-check and fallback-source layers; the migration logic that unconditionally rewrites provider: "auto" to "openai" is not changed and requires maintainer product decision per #90787's clawsweeper:needs-product-decision label.

Verification test output (12 tests, including 3 new):

 RUN  v4.1.7

 ✓ |extension-memory| extensions/memory-core/src/memory/manager-reindex-state.test.ts (12 tests) 31ms

 Test Files  1 passed (1)
      Tests  12 passed (12)

Key new test assertions:
- "falls back to fts-only when provider.model is an empty string":
    provider.model = "" → expectedModel = "fts-only" → matches fts-only index → status: "valid"
- "reports mismatch when empty-string expected model is compared to a non-fts index":
    provider.model = "" → expectedModel = "fts-only" → mismatched with "text-embedding-3-small" → reason contains "expected fts-only"
- "falls back to fts-only when provider.model is whitespace-only":
    provider.model = "  " → trimmed to "" → expectedModel = "fts-only" → matches fts-only index → status: "valid"

Risks

Low risk: Two-point guard — one at the identity comparison consumer, one at the fallback-source parameter — preserving existing behavior for all non-empty provider.model values.
Additive: The guards only activate when provider.model is empty or whitespace, or when fallbackSourceModel would otherwise be "". Previously these produced broken empty-string mismatches. No change for correctly-populated provider.model.
Complementary with Gateway memory_search index stuck dirty: provider.model empty during boot overwrites correct index identity #90042: PR Gateway memory_search index stuck dirty: provider.model empty during boot overwrites correct index identity #90042 fixes provider.model at the creation source (createWithAdapter); this PR protects the identity check at the consumption layer and the fallback-source computation. Either fix alone reduces the bug surface; together they eliminate it.
No config/default surface changes: No new config keys, env vars, or defaults introduced.
Adapter-default resolution preserved: resolveEmbeddingProviderFallbackModel("openai", "fts-only", cfg) still returns "text-embedding-3-small" (adapter defaultModel) when the adapter is registered. "fts-only" only appears when the adapter is unregistered — which is the correct semantic for the no-provider path.

clawsweeper · 2026-06-09T12:21:19Z

Codex review: needs maintainer review before merge. Reviewed June 15, 2026, 5:10 AM ET / 09:10 UTC.

Summary
This PR changes memory-core index identity resolution so empty or whitespace provider models and unresolved configured-provider fallback models no longer produce an empty expected model, and adds focused regression tests.

PR surface: Source 0, Tests +35. Total +35 across 3 files.

Reproducibility: yes. at source level: current main still reads raw provider.model in the identity comparator and uses an empty fallback source in the configured-provider path. I did not run a full packaged upgrade reproduction in this read-only review.

Review metrics: 1 noteworthy metric.

Identity Fallback Points: 2 changed. Both the direct identity comparator and the pre-init configured-provider fallback influence upgrade dirty-state and vector-search availability.

Stored data model
Persistent data-model change detected: vector/embedding metadata: extensions/memory-core/src/memory/manager-reindex-state.test.ts. Confirm migration or upgrade compatibility proof before merge.

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] Get memory-owner confirmation for the fts-only unresolved-provider fallback before merge.

Risk before merge

[P1] The PR intentionally changes how empty or unresolved memory provider models compare against persisted index metadata; that can alter dirty, reindex, and vector-search availability for upgraded users.
[P1] If maintainers expect an initialized provider with an empty model to resolve the adapter default or configured model instead of comparing as fts-only, the fix should move to the provider model source-of-truth or caller normalization before merge.

Maintainer options:

Confirm Memory Fallback Semantics (recommended)
Have a memory owner confirm that unresolved configured providers should compare as fts-only before this lands.
Normalize At Provider Source
If initialized empty provider models should use adapter defaults, repair the provider creation or caller normalization layer instead of treating them as fts-only.
Pause For Broader Migration Policy
If this is inseparable from the legacy auto/Gemini upgrade behavior, pause this PR and keep that policy decision on [Bug]: memorySearch provider silently resets to "openai" after upgrade to 2026.6.1, causing permanent Dirty index and vector search outage #90787.

Next step before merge

[P2] Maintainer review should decide the fallback semantics; the previous mechanical branch blocker appears removed and there is no narrow automated repair to apply now.

Security
Cleared: The diff touches memory identity fallback logic and tests only; it does not change dependencies, workflows, secrets, install scripts, package metadata, or other supply-chain surfaces.

Review details

Best possible solution:

Land the narrow empty-expected guard only after memory owners confirm the fts-only fallback semantics, or adjust the branch to resolve/backfill the canonical provider model before identity comparison.

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

Yes at source level: current main still reads raw provider.model in the identity comparator and uses an empty fallback source in the configured-provider path. I did not run a full packaged upgrade reproduction in this read-only review.

Is this the best way to solve the issue?

Not fully settled. The PR is a plausible narrow mitigation, but the best merge state needs owner confirmation that unresolved configured providers should compare as fts-only rather than resolving or backfilling the canonical provider model.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 501f63443f54.

Label changes

Label changes:

add proof: sufficient: Contributor real behavior proof is sufficient. The PR body supplies after-fix terminal output from focused runtime scripts and targeted test output exercising the empty-model identity and fallback paths.
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 (terminal): The PR body supplies after-fix terminal output from focused runtime scripts and targeted test output exercising the empty-model identity and fallback paths.
remove rating: 🦐 gold shrimp: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.
remove status: ⏳ waiting on author: Current PR status label is status: 👀 ready for maintainer look.

Label justifications:

P2: The PR targets a normal-priority memory identity regression with limited scope but real vector-search impact for affected configurations.
merge-risk: 🚨 compatibility: The fallback changes how existing empty or unresolved provider model state is interpreted across upgrades and status checks.
merge-risk: 🚨 session-state: Memory index identity controls dirty, reindex, and vector-search availability state for persisted memory/session recall.
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 (terminal): The PR body supplies after-fix terminal output from focused runtime scripts and targeted test output exercising the empty-model identity and fallback paths.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body supplies after-fix terminal output from focused runtime scripts and targeted test output exercising the empty-model identity and fallback paths.

Evidence reviewed

PR surface:

Source 0, Tests +35. Total +35 across 3 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	2	2	2	0
Tests	1	35	0	+35
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	3	37	2	+35

What I checked:

Current PR surface: Live PR metadata shows the current head only changes the memory reindex-state test, memory identity comparator, and memory sync identity fallback files; the earlier src/scripts/test-projects.test.ts route expectation is no longer in the PR file list. (1dd9bccdb2a8)
Current main identity behavior: Current main computes expectedModel directly from params.provider.model, so an empty provider model can still produce an empty expected model before this PR. (extensions/memory-core/src/memory/manager-reindex-state.ts:162, 501f63443f54)
Current main configured-provider fallback: Current main passes an empty string as the fallback source model for configured-provider status checks before provider initialization; the PR changes this to fts-only. (extensions/memory-core/src/memory/manager-sync-ops.ts:520, 501f63443f54)
Provider default contract: The OpenAI memory adapter declares text-embedding-3-small as its default model and the Google/Gemini memory adapter declares gemini-embedding-001, so provider-default resolution is a real identity source of truth when the adapter is registered. (extensions/openai/memory-embedding-adapter.ts:16, 501f63443f54)
Docs compatibility contract: The memory config docs say provider/model identity changes pause vector search until rebuild, which makes this PR's fallback behavior compatibility- and session-state-sensitive even with green CI. Public docs: docs/reference/memory-config.md. (docs/reference/memory-config.md:61, 501f63443f54)
Related reports: The related reports describe empty expected-model memory identity failures and broader provider migration concerns; [Bug]: memorySearch provider silently resets to "openai" after upgrade to 2026.6.1, causing permanent Dirty index and vector search outage #90787 remains open with a product-decision label, so this PR should not claim the whole upgrade policy is settled.

Likely related people:

mushuiyu886: Authored the recent merged memory identity compatibility work in the same reindex-state and sync-ops paths. (role: recent area contributor; confidence: high; commits: 44e6caff5401; files: extensions/memory-core/src/memory/manager-reindex-state.ts, extensions/memory-core/src/memory/manager-sync-ops.ts)
jalehman: Current blame for the extracted identity comparator and configured-provider fallback scaffold points to the bundled plugin session seam refactor. (role: recent current-file contributor; confidence: medium; commits: f1b8827d20c8; files: extensions/memory-core/src/memory/manager-reindex-state.ts, extensions/memory-core/src/memory/manager-sync-ops.ts)
vincentkoc: Recent related memory identity and recovery reviews/merges connect this person to the same memory-core identity and reindex safety surface. (role: adjacent memory contributor; confidence: medium; commits: 44e6caff5401, d46dc39b18ec; files: extensions/memory-core/src/memory/manager-reindex-state.ts, extensions/memory-core/src/memory/manager-sync-ops.ts, extensions/memory-core/src/memory/manager.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.

xydt-tanshanshan · 2026-06-09T12:44:04Z

@clawsweeper re-review

clawsweeper · 2026-06-09T12:44:08Z

🦞🧹
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: Failed
Detail: The targeted re-review did not finish cleanly. Check the workflow run for details.
Run: https://github.com/openclaw/clawsweeper/actions/runs/27391463051
Updated: 2026-06-12T03:00:18.865Z

xydt-tanshanshan · 2026-06-10T00:45:29Z

@clawsweeper re-review

clawsweeper · 2026-06-10T00:46:25Z

🦞👀
ClawSweeper picked this up.

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

xydt-tanshanshan · 2026-06-11T07:33:53Z

@clawsweeper re-review

clawsweeper · 2026-06-11T07:33:56Z

🦞👀
ClawSweeper picked this up.

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

Re-review progress:

State: Failed
Detail: The targeted re-review did not finish cleanly. Check the workflow run for details.
Run: https://github.com/openclaw/clawsweeper/actions/runs/27331376303
Updated: 2026-06-11T07:39:32.302Z

xydt-tanshanshan · 2026-06-12T02:54:14Z

@clawsweeper re-review

…IndexIdentityState Refs openclaw#90787

xydt-tanshanshan · 2026-06-15T08:37:12Z

@clawsweeper re-review

clawsweeper · 2026-06-15T08:37:15Z

🦞🧹
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: Superseded
Detail: A newer re-review for this item started before this run finished, so GitHub cancelled this older run. Check the latest ClawSweeper run for the current result.
Run: https://github.com/openclaw/clawsweeper/actions/runs/27534311868
Updated: 2026-06-15T08:40:28.470Z

…ead of empty string Change resolveEmbeddingProviderFallbackModel fallbackSourceModel from empty string to fts-only so unregistered adapters produce a readable expectedModel instead of an empty string that causes permanent identity mismatch. Refs openclaw#90787

xydt-tanshanshan · 2026-06-15T09:00:57Z

@clawsweeper re-review

clawsweeper · 2026-06-15T09:01:00Z

🦞🧹
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/27535503689
Updated: 2026-06-15T09:11:41.302Z

Prove that existing indexes with empty adapter model metadata transition safely after the backfill: stored meta (written by the identity path) already carries the correct model; after upgrade, provider.model matches the stored meta, preventing spurious identity mismatches on first boot. Ref: openclaw#90042, openclaw#91691

openclaw-barnacle Bot added extensions: memory-core Extension: memory-core size: XS proof: supplied External PR includes structured after-fix real behavior proof. labels Jun 9, 2026

clawsweeper Bot added the merge-risk: 🚨 compatibility 🚨 May break existing users, config, migrations, defaults, or upgrade paths. label Jun 9, 2026

[AI] fix(memory): prevent empty-string expectedModel in resolveMemory…

f557376

…IndexIdentityState Refs openclaw#90787

xydt-tanshanshan force-pushed the fix/memory-identity-empty-model-90787 branch from 64feeb5 to e83e1e5 Compare June 15, 2026 08:31

xydt-tanshanshan force-pushed the fix/memory-identity-empty-model-90787 branch from e83e1e5 to 878bcd8 Compare June 15, 2026 08:40

xydt-tanshanshan force-pushed the fix/memory-identity-empty-model-90787 branch from 878bcd8 to f30016d Compare June 15, 2026 08:51

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label Jun 15, 2026

xydt-tanshanshan force-pushed the fix/memory-identity-empty-model-90787 branch from f30016d to 1dd9bcc Compare June 15, 2026 08:54

This was referenced Jun 15, 2026

Gateway memory_search index stuck dirty: provider.model empty during boot overwrites correct index identity #90042

Open

[Bug]: memory status shows inconsistent Dirty state, watcher stops auto-indexing, and --deep required to confirm health on 2026.6.1 #90801

Open

vincentkoc self-assigned this Jun 15, 2026

vincentkoc merged commit 85a6353 into openclaw:main Jun 15, 2026
178 of 183 checks passed

Uh oh!

Uh oh!

Conversation

xydt-tanshanshan commented Jun 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Change Type

Scope

Linked Issue

Motivation

Changes

Verification

Real Behavior Proof

Risks

Uh oh!

clawsweeper Bot commented Jun 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

xydt-tanshanshan commented Jun 9, 2026

Uh oh!

clawsweeper Bot commented Jun 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

xydt-tanshanshan commented Jun 10, 2026

Uh oh!

clawsweeper Bot commented Jun 10, 2026

Uh oh!

xydt-tanshanshan commented Jun 11, 2026

Uh oh!

clawsweeper Bot commented Jun 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

xydt-tanshanshan commented Jun 12, 2026

Uh oh!

xydt-tanshanshan 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!

xydt-tanshanshan 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!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

xydt-tanshanshan commented Jun 9, 2026 •

edited

Loading

clawsweeper Bot commented Jun 9, 2026 •

edited

Loading

clawsweeper Bot commented Jun 9, 2026 •

edited

Loading

clawsweeper Bot commented Jun 11, 2026 •

edited

Loading

clawsweeper Bot commented Jun 15, 2026 •

edited

Loading

clawsweeper Bot commented Jun 15, 2026 •

edited

Loading