Repository navigation
fix: assert DingTalk empty-config Gateway responses - #477
Conversation
|
@clawsweeper review |
|
🦞👀 Re-review progress:
|
|
Codex review: needs maintainer review before merge. What this changesThis PR declares DingTalk’s twelve empty-config Gateway rejection contracts and forwards exact expectations through both synthetic execution routes. Example: An empty-config probe invokes dingtalk-connector.sendToUser.
Review scores
ProductKind: Feature · Worth it: Yes · Fix scope: Complete Merge readiness✅ Ready for maintainer review Keep open: this maintainer-authored PR adds useful fixture assertions absent from main, and review found no concrete introduced defect. Priority: P2 Before mergeNone. FindingsNone. Tests
Agent review detailsHow this fits togetherCrabpot supplies fixture declarations and retained plugin handlers to Plugin Inspector, which executes synthetic probes and returns pass, fail, or blocked results. flowchart LR
A[Fixture manifest] --> B[Crabpot synthetic runner]
C[Captured or entrypoint handlers] --> B
B --> D[Plugin Inspector]
D --> E[Exact response comparison]
E --> F[Probe report]
Technical reviewBest possible solution: Keep exact response comparison in Inspector and fixture-owned declarations in Crabpot while preserving undeclared failures and separate dependency qualification. Do we have a high-confidence way to reproduce the issue? Source inspection establishes that the retained DingTalk handlers intentionally reject empty requests while main lacks fixture declarations to assert those responses; no target code was executed during this read-only review. Is this the best way to solve the issue? The implementation uses the existing response owner and fixture-selection path, providing exact negative-contract coverage without skipping handlers or broadly suppressing rejections. AGENTS.md: found and applied where relevant. Codex review notes: model internal, reasoning medium; reviewed against d300f1bf92fb. Provenance checked
TestingProof path: in-process harness. SecurityNone. EvidenceWhat I checked:
Likely related people:
Review metrics
LabelsLabel changes:
Label justifications:
Rank-up movesOptional improvements that raise the rating; they are not merge blockers.
Rating scale6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior. WorkflowClawSweeper edits this one comment on every review. Comment Reviewed October 14, 2026, 11:26 PM ET / October 15, 2026, 03:26 UTC. |
|
ClawSweeper reviews are opt-in for maintainers. If a review is required, finish the work and validation first, then comment |
What Problem This Solves
Crabpot cannot declare exact method-specific Gateway responses, so intentional empty-config rejections appear as unexpected failures.
User Impact
Fixture-owned expectations execute each retained handler and require its exact response. Undeclared rejections, mismatches, missing methods, blocked expected methods, and handler failures remain failures. Both captured and entrypoint execution routes use the same declarations.
Why This Change Was Made
Validate the optional expectation map in the manifest and schema, forward it through both synthetic routes, and declare DingTalk's twelve existing credential-free rejection contracts. Adopt landed Inspector source
9df499d8884883bfaaeb1eeaac0da5ffe9b9e80bfrom openclaw/plugin-inspector#110. The published CLI remains pinned to0.3.26, which does not contain the new API; source and locally packed delivery provide it without publication.Adds 35 net runtime lines, 25 schema lines, 16 fixture lines and 67 net test lines. No submodule or package version changes are included.
Evidence
a61680cbdd46390be713df7bd2da11eea7ddafd2received a fresh P0–P2 automated review after adopting the landed source pin.9468d9a33927720de29dacd1c4c56aad62548315is byte-identical to the reviewed and executed source head955863f12ec3f552cbeb8a7d7b9d03d7ab93b7e4.run_20ebfd050066used Node 24.19.0, Crabpotb9f6fcd06a1dead9dbac567f696e463fccbeb740and that Inspector source. Consumer tests passed 68/68, including both execution routes and source-loader cases; Inspector focused/sibling tests passed 98/98. All 61 unpublished package files matched source.5fef12d37377e299e26d18b0145baf646cd17a8b(0.8.26) with its declared OpenClaw 2026.8.1 dependency, then exercised it under the mock SDK. Baseline Inspector produced 3 pass / 12 fail / 1 blocked; source and packed candidates each produced 15 pass / 0 fail / 1 blocked. It verified twelve exact negative responses and three exact diagnostics. A deliberate mismatch produced 14 pass / 1 fail / 1 blocked; the channel prerequisite remained blocked.DingTalk has no tracked lockfile. That run generated one and installed with
npm ci; lock SHA256:22e2d47f5f7ff6e4888c2b8b93edc9aec9868258b8e0a95e45451bd6a5bef8f8. This is qualification of that resolved build and the assertion path under the mock SDK, not live DingTalk or full host compatibility. The prior run is not a fresh execution of this final commit; final CI checks the landed source-pin integration. The initial local source-loader attempt failed three cases before the Inspector commit was pushed; the isolated run passed those cases.Related: #327 remains open and held. This PR retains the current DingTalk gitlink
39bdb2d8a7468237858906d255a717cd03b037ad; it does not adopt5fef12d37377e299e26d18b0145baf646cd17a8b, supersede the dependency PR, or claim its upgrade is ready. The installed mandatory review helper cannot admit a gitlink diff, so that dependency change is outside this PR's explicit scope.Fixture Impact
Notes
No credentials, network-backed plugin operation, package publication, tag, deployment or release is included. Published CLI remains pinned to 0.3.26.
Final Review Follow-through
The fresh ClawSweeper review covers
a61680cbdd46390be713df7bd2da11eea7ddafd2: no findings or before-merge requirements. It independently verified the twelve guards at the retained DingTalk revision.The optional rank-up suggestion to link raw output is acknowledged. The existing proof summary and private-log decision apply: coordinator and executor directly inspected the source, unpublished packed, and deliberate-mismatch results. Raw lease logs remain private; no new artifact upload or repeat run is needed. This does not extend the reported target-revision proof to a fresh execution of the retained revision.
Fresh CI passed Ubuntu, macOS, Windows, and container static checks at this head. The Ubuntu source-and-packed Inspector smoke, runtime smoke, and resource collector checks passed. The changed-fixture lane is correctly unselected because no submodule pin changes.