| name | storybook-review |
| description | Use when asked to review Storybook stories or Storybook testing coverage in this repository, especially `*.stories.*` files. Apply `.agents/skills/storybook/SKILL.md` as the rubric and write the review to `.review/storybook.md`. |
Storybook Review
Read storybook/SKILL.md first. Treat every rule in that skill as the review rubric.
Do not modify the reviewed Storybook files while reviewing unless the user explicitly asks for fixes.
Write the final review to .review/storybook.md. Create .review/ if it does not exist. Overwrite the file on each fresh review so it only contains the current result.
Capture the current git SHA before the review begins and preserve that exact SHA in the report.
Prefer file citations as path:line. If exact line numbers are unavailable, cite the file path.
Use these review sections:
Coverage
Play Tests
Reuse
Browser Environment
Section rules:
Status is PASS only when every reviewed file satisfies that section.
Status is FAIL when any reviewed file has one or more unresolved issues in that section.
Evidence must cite the relevant files and explain the observed behavior.
Recommendations must be exactly None on PASS.
Recommendations must be concrete, imperative remediations on FAIL.
Overall rules:
Overall is PASS only if every section is PASS. Otherwise FAIL.
- Generate
## Action checklist from the Recommendations text of every failing section, in section order.
- If a failing section needs multiple concrete fixes, split them into multiple checklist items.
- Keep checklist items specific and actionable. Name the missing state, the missing or weak
play test, the duplicated setup that should be extracted, the hard-coded timeout that must be removed, or the browser-environment violation that must be corrected.
- Include the affected file path in each checklist item when more than one file is reviewed.
- Do not generate checklist items for passing sections.
- If every section passes, write exactly
- [x] No action items.
- Keep the report factual. Do not add praise, scoring, or optional suggestions outside the required format.
- Return Markdown only.
Use this exact report shape:
# Storybook Review
Git SHA: <exact git SHA captured before the review>
Overall: <PASS|FAIL>
Files reviewed: <comma-separated list of reviewed files>
Reference rubric: `.agents/skills/storybook/SKILL.md`
Summary: <short summary>
## Action checklist
- [ ] <one checklist item per concrete remediation, in failing-section order; if none, write exactly `- [x] No action items.`>
## Coverage
Status: <PASS|FAIL>
Evidence: <file-specific evidence with citations>
Recommendations: <None or concrete remediation>
## Play Tests
Status: <PASS|FAIL>
Evidence: <file-specific evidence with citations>
Recommendations: <None or concrete remediation>
## Reuse
Status: <PASS|FAIL>
Evidence: <file-specific evidence with citations>
Recommendations: <None or concrete remediation>
## Browser Environment
Status: <PASS|FAIL>
Evidence: <file-specific evidence with citations>
Recommendations: <None or concrete remediation>