メニュー
Audits a Meridian user story against writing quality, section contracts and refine checklist — report only, no ready flag. Use for /review-us US-XXXX, quality audit, or before refine/implement.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Closes a Meridian sprint after sprint review — fills Retrospective, sets status complete, logs decisions. Use with /complete-sprint vX-SY.
Closes a Meridian user story after implementation — fills Record, acceptance, status and board. Use when marking US done, completing US-XXXX, or after implementing a user story.
Creates a Meridian epic file in docs/epics after architecture is approved. Use when defining a new product capability block before user stories.
Creates a Meridian sprint file in docs/sprints linked to a version. Use when planning execution slices within a release.
Creates a valid Meridian user story after epics and versions are approved. Use when adding work to docs/us and keeping acceptance criteria concrete.
Creates a Meridian release file in docs/versions. Use when defining a new product version before user stories.
| name | review-user-story |
| description | Audits a Meridian user story against writing quality, section contracts and refine checklist — report only, no ready flag. Use for /review-us US-XXXX, quality audit, or before refine/implement. |
| allowed-tools | Read, Glob, Grep, Bash |
Read-only audit. Compare the US file to templates and checklists; output a gap report. Do not set
ready: true, edit the US, or write product code unless the manager explicitly asks to fix in the same turn.
| File | When to read |
|---|---|
.agent/references/templates/TEMPLATE_SOURCES.md | When unsure which file is canonical |
.agent/references/templates/writing-guide.md | Prose quality bar (Why / Where / Approach) |
.agent/references/templates/section-contracts.md | Fixed ## / ### structure |
references/review-checklist.md | Mandatory — audit rubric and output format |
references/us-template.md | Expected full US shape |
.agent/skills/refine-user-story/references/refine-checklist.md | Same gates /refine-us uses — cite failures here |
Target docs/us/US-XXXX.md | File under review |
/create-us, before /refine-us — “is this draft good enough to refine?”ready: true/review-us US-XXXX/refine-us/review-us | /refine-us | |
|---|---|---|
| Edits US | No (default) | Yes — deepens Context |
Sets ready: true | Never | When checklist passes |
| Product code | Never | Never |
| Output | Gap report + recommendation | Updated US file |
If review finds blockers → recommend /refine-us US-XXXX. If review passes all refine gates → say “ready for /refine-us to set ready: true” (refine still required for the flag).
depends_on US if cited).python3 .agent/scripts/validate_meridian.py <project-folder> when available; include errors/warnings in report.writing-guide.md anti-patterns.ready, status, depends_on, tests / tests_status./refine-us, /create-us rework, implement (only if ready + passes), or /complete-us (if code done).US review:
File:
Validator: pass | N errors, M warnings
Checklist: X/Y pass
Failures:
- [check id] — what is wrong — where in file
Warnings:
- ...
Ready for implement (content): yes | no
ready flag in file: true | false | unset
Recommendation: /refine-us US-XXXX | implement | /complete-us | human edit