ci: use a seeded fixture for skill hero layout checks

[steipete]

The public browser smoke gate is red because its mobile layout test hard-codes a third-party skills.sh listing that now returns HTTP 404. This fails unrelated PRs before any layout assertion runs.

Move the existing layout assertions into the local-auth profile/context shard and seed the existing doany-skills/skills/reddit-automation fixture before navigation. The test keeps its desktop/mobile typography, grouping, wrapping, and overflow checks. Its topic-count assertion accounts for the seeded entry’s two original topics. Product UI behavior is unchanged.

Evidence: failing main run, with only this test failing (18 other smoke tests passed); the hard-coded public route independently returned HTTP 404.

Remote proof passed on the configured Blacksmith Testbox backend, lease tbx_01m33nwh6z4rcm0er9wakvcwkd (backend run):

• bun run test:pw:local-auth -- --project=chromium e2e/local-auth/skill-hero-layout.pw.test.ts: real disposable Convex and Chromium, 1 passed.
• bun run ci:playwright-smoke: 18 passed.
• bun run ci:static and bun run ci:unit: passed; 7,170 tests passed, 3 skipped.
• Schema and CLI no-emit TypeScript checks: passed.
• Independent Codex autoreview: scoped-clean through P2.

AI-assisted maintenance fix; the code and runtime proof have been reviewed. No release or production deployment is requested.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
clawhub	Ready	Preview	Sep 22, 2026 5:06am UTC

[clawsweeper]

Codex review: blocked before merge.

What this changes

The PR moves skill-page layout checks from the public catalog to a seeded local browser test, preventing removed third-party listings from failing unrelated changes.

Review scores

Measure	Result	What it means
Overall readiness	🦐 gold shrimp (3/6)	The scoped patch has convincing browser execution evidence, but its additional product value is no longer established.
Proof confidence	🦞 diamond lobster (5/6)	Sufficient (live_output): The complete body records the changed fixture-backed layout spec passing through the real local app, disposable Convex backend, and Chromium; exact-head profile-context success corroborates it. Raw job logs were blocked by the review proxy. No stored-data contract changes.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Product

Kind: Maintenance · Worth it: No · Fix scope: Complete
User problem: Unrelated contributions fail browser CI when the particular public skill used by a layout test disappears.
Reason: The original fixture migration was useful CI maintenance, but current main now provides equivalent coverage and no distinct remaining benefit was demonstrated.

Merge readiness

⛔ Blocked before merge - 2 items remain

Current main covers the central problem through a later merged repair, but the required formal fixing link for automated PR closure is absent. No introduced code defect was found.

Priority: P3
Reviewed head: 6e343c3f7f4c9d451e0fcee3b443554043dca825

Before merge

• Complete next step - Resolve the GitHub-reported merge conflicts against current main if continuing this PR.
• Product: not worth merging - The original fixture migration was useful CI maintenance, but current main now provides equivalent coverage and no distinct remaining benefit was demonstrated.

Findings

None.

Agent review details

How this fits together

ClawHub browser CI checks rendered catalog pages for regressions. Its local browser runner starts an isolated app and Convex backend, supplies fixture data, and reports browser assertion results.

flowchart LR
  A[Pull request CI] --> B[Local browser shard]
  B --> C[Isolated app and backend]
  D[Seeded catalog entry] --> C
  C --> E[Chromium skill page]
  E --> F[Typography and wrapping checks]
  F --> G[CI result]

Loading

Technical review

Best possible solution:

Keep main's shared fixture initialization and existing metadata spec, concentrating any demonstrated coverage gap there.

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

Not applicable to a current product bug: source inspection confirms main has removed the unstable public-listing dependency and retained the layout assertions.

Is this the best way to solve the issue?

No additional implementation is currently justified: main already seeds the fixture through the shared runner and exercises equivalent assertions in the required browser shard.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against d044664a7636.

Provenance checked

• e2e/public-routes-smoke.pw.test.ts metadata layout assertions keeps the original intent (fix: preserve semantic skill metadata wrapping #3735: Check semantic metadata wrapping, attached dividers, secondary typography, and actual geometry while preserving plugin layout.)
• .github/workflows/ci.yml local browser shard routing keeps the original intent (fix(ci): bundle ClawHub PR gates #2838: Group browser specs to reduce runner registrations while preserving required coverage.)
• specs/ci.md local browser gate keeps the original intent (test: add skill publish lifecycle e2e #2193: Provide isolated, contributor-runnable browser coverage using local Convex and dev authentication.)

Testing

Proof path: shipped entry point.

Security

None.

Evidence

What I checked:

• Pinned introduced changes: The verified merge-base-to-head diff moves the existing browser assertions, seeds the reddit-automation fixture, adjusts topic counting, and registers the test in profile-context. No application production code changes. (e2e/local-auth/skill-hero-layout.pw.test.ts:10, 6e343c3f7f4c)
• Equivalent current-main coverage: Main's metadata test uses the same local reddit-automation route and retains desktop typography, mobile grouping, wrapping, overflow, and runtime-error assertions. It normalizes the stress setup to three category and topic nodes; no distinct coverage gap from the prior rank-up move was established. (e2e/local-auth/skill-hero-metadata.pw.test.ts:9, d044664a7636)
• Shared fixture initialization: Current main seeds the canonical fixture before building the app and includes the metadata test in the required profile-context shard. Public smoke no longer contains the unstable skill layout test. (scripts/run-playwright-local-auth.ts:693, d044664a7636)
• Merged overlapping repair: fix(workers): preserve explicit scanner model settings #3802 is verified merged and explicitly describes replacing this same live-listing dependency with controlled local coverage. Its body does not formally identify this PR as the request it fixes. (9a614dc195af)
• Fixture migration history: Scoped feature history identifies the merged repair as the addition of fixture-backed metadata coverage; no later main commit changes that spec. (e2e/local-auth/skill-hero-metadata.pw.test.ts:9, 9a614dc195af)
• Original layout intent: fix: preserve semantic skill metadata wrapping #3735 specifies semantic wrapping, attached dividers, secondary typography, and geometry-based browser checks. The inspected commit patch adds the original regression; this branch preserves those assertions. (e2e/public-routes-smoke.pw.test.ts:234, c1c9cfbd04c2)

Review metrics

Metric	Value	Why it matters
Production and test LOC	production +0; browser tests +181/-163; workflow +1; docs +4	The branch relocates browser coverage without growing application production code.

Root-cause cluster

Relationship: superseded
Canonical: #3802
Summary: The merged fixture migration covers the same browser CI failure, without establishing the formal fixing link required for automated PR closure.

Members:

• canonical: fix(workers): preserve explicit scanner model settings #3802 - Its merged changes move the same assertions to controlled fixture data and the required local browser shard.

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Labels

Label changes:

No label changes.

Label justifications:

• P3: This is browser CI maintenance whose central reliability repair already exists on main.
• rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦞 diamond lobster and patch quality is 🐚 platinum hermit.
• status: ⏳ waiting on author: ClawSweeper has contributor-facing work open and is waiting for author action.
• proof: sufficient: Contributor real behavior proof is sufficient.

Rating scale

6/6 🦀 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.

Workflow

ClawSweeper edits this one comment on every review. Comment @clawsweeper re-review for a fresh review only; repair and merge need explicit maintainer commands such as @clawsweeper autofix or @clawsweeper automerge.

History

Review history (19 earlier review cycles; latest 8 shown)

• reviewed 2026-10-06T18:43:01.503Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-07T06:53:10.662Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-07T12:49:51.251Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-07T17:57:08.314Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-08T03:44:51.892Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-08T15:38:14.110Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-08T21:37:01.097Z sha 6e343c3 :: blocked before merge. :: none
• reviewed 2026-10-09T04:43:20.253Z sha 6e343c3 :: blocked before merge. :: none

Reviewed October 9, 2026, 7:34 AM ET / 11:34 UTC (Revision 20).