Content
90%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A lean, highly actionable skill body that assumes Claude's competence while supplying only repo-specific knowledge, with a conservative dry-run-first workflow appropriate to its destructive nature. The only structural improvement would be a post-run verification step for the diary edit and slightly tighter reference navigation.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every section carries project-specific facts Claude cannot know (the AGENTS.md lifecycle rule, the diary update protocol, env var defaults and their rationale) with no generic git or worktree concept explanations and no padding — every token earns its place. | 5 / 5 |
Actionability | Copy-paste-ready commands ('sh scripts/worktree-cleanup.sh --dry-run', 'sh scripts/worktree-cleanup.sh'), documented env vars with defaults, and exact diary-edit semantics including the 'already stale' edge case fully cover the common cases. | 5 / 5 |
Workflow Clarity | For a destructive/batch skill, validation is present — a mandatory dry-run shown to the operator, conservative script behavior, an explicit dirty-root error path, and a report step — so the score-3 cap does not apply; however there is no post-run verification of the diary edit, leaving a minor checkpoint gap between anchors 4 and 5. | 4 / 5 |
Progressive Disclosure | Well-organized sections with git mechanics correctly delegated to scripts/worktree-cleanup.sh and all cross-references (AGENTS.md, docs/diary.md, /merge-train, constitution/local-workflow.md) one level deep and clearly signaled; at ~75 lines with all instructional content inline rather than split into navigable reference files, it sits just below the anchor-5 structure. | 4 / 5 |
Total | 18 / 20 Passed |