test: link ClawHub packaged plugin QA evidence#97712
Conversation
|
Codex review: needs maintainer review before merge. Reviewed June 29, 2026, 1:40 PM ET / 17:40 UTC. Summary PR surface: Other +64. Total +64 across 3 files. Reproducibility: not applicable. this is QA metadata wiring rather than a bug report. The review path is source validation of scenario discovery, taxonomy IDs, referenced proof paths, and the latest PR diff. Review metrics: 1 noteworthy metric.
Stored data model Merge readiness Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch. Rank-up moves:
Risk before merge
Maintainer options:
Next step before merge
Security Review detailsBest possible solution: Land the narrow QA metadata after the PR body reflects the latest two-scenario diff and exact-head checks finish, keeping the existing proof scripts as the execution source instead of duplicating assertions in new scenario logic. Do we have a high-confidence way to reproduce the issue? Not applicable; this is QA metadata wiring rather than a bug report. The review path is source validation of scenario discovery, taxonomy IDs, referenced proof paths, and the latest PR diff. Is this the best way to solve the issue? Yes for the patch shape: current QA Lab contracts already support nested YAML, vitest execution, and script execution with args, and the maturity investigation asks for small wrappers around existing proof. The PR body should be refreshed before merge because it no longer matches the latest two-scenario diff. AGENTS.md: found and applied where relevant. Codex review notes: model internal, reasoning high; reviewed against 597a0ba43ca3. Label changesLabel justifications:
Evidence reviewedPR surface: Other +64. Total +64 across 3 files. View PR surface stats
What I checked:
Likely related people:
What the crustacean ranks mean
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
|
* test: link ClawHub packaged plugin QA evidence * test: narrow ClawHub QA primary links
* test: link ClawHub packaged plugin QA evidence * test: narrow ClawHub QA primary links
* test: link ClawHub packaged plugin QA evidence * test: narrow ClawHub QA primary links
* test: link ClawHub packaged plugin QA evidence * test: narrow ClawHub QA primary links
What Problem This Solves
The QA maturity inventory still had existing ClawHub and packaged-plugin proof that was not represented as native QA scenario evidence. That left these existing coverage IDs under-counted even though the repo already had targeted release/package proof:
clawhub.marketplace-listclawhub.npm-pack-local-release-candidate-installsplugins.packaged-bundled-pluginsWhy This Change Was Made
This adds three QA scenario YAML entries that link the existing proof into QA Lab metadata without changing taxonomy IDs. The release marketplace proof stays on its existing Docker release lane through
scripts/qa/docker-e2e-lane.ts; the release-candidate checklist and packaged bundled plugin coverage use their focused Vitest proof instead of broad or unsharded Docker sweeps.The underlying scripts stay in
scripts/because these are active release, package, Docker scheduler, or plugin-prerelease pathways rather than orphaned proof scripts.User Impact
Maintainers get primary QA coverage accounting for the existing ClawHub marketplace, release-candidate package install checklist, and packaged bundled plugin proof. No runtime behavior changes.
Definite primary coverage added:
clawhub.marketplace-list:clawhub-marketplace-listruns the package-installed release plugin marketplace lane, which creates a fixture marketplace, lists it as JSON, installs by marketplace shortcut, updates by plugin id, and uninstalls.clawhub.npm-pack-local-release-candidate-installs:clawhub-release-candidate-checklistpoints at the release checklist tests that carry the exact prepared candidate tarball into npm-update install proof and validate npm preflight tarball metadata/digest before publish.plugins.packaged-bundled-plugins:packaged-bundled-plugin-install-uninstallpoints at the bundled plugin probe tests that select package-root bundled plugins, reject source-only roots, and assert packaged install/uninstall record paths.Evidence
pnpm exec oxfmt --write scripts/qa/docker-e2e-lane.tsnode scripts/run-vitest.mjs extensions/qa-lab/src/scenario-catalog.test.ts test/scripts/release-candidate-checklist.test.ts test/scripts/bundled-plugin-install-uninstall-probe.test.ts --reporter=verbosepnpm openclaw qa coverage --match 'clawhub marketplace' --jsonpnpm openclaw qa coverage --match 'packaged bundled plugin' --jsonpnpm openclaw qa coverage --match 'release candidate checklist' --jsonpnpm openclaw qa coverage --jsonreportsscenarios=153 coverageIds=204 primary=134 secondary=98and primary rows for all three target IDs.bash -n scripts/e2e/release-plugin-marketplace-docker.sh scripts/e2e/bundled-plugin-install-uninstall-docker.sh scripts/e2e/lib/bundled-plugin-install-uninstall/sweep.sh scripts/e2e/lib/release-plugin-marketplace/scenario.sh.agents/skills/autoreview/scripts/autoreview --mode localclean after fixing the initial unsharded bundled-plugin lane concern.Heavy Docker lanes were not run locally for this metadata PR; the PR links existing release/package proof and validates the native QA inventory plus focused tests.