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.
An exemplary instruction-only skill body: fully executable commands, app-specific knowledge only, and a safety-conscious preview/confirm/apply pattern for destructive operations. The remaining gaps are a missing post-apply verification step and modest room to offload the gotchas into a reference file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every section is lean and app-specific — retention model, pinned-audio behavior, size ratios, and the iCloud caveat are things Claude cannot know — with no generic explanations or padding; commands carry brief, purposeful comments. | 5 / 5 |
Actionability | Every task ships copy-paste-ready commands ("minutes storage --json", "minutes cleanup --older-than 14d --apply", "find ~/meetings -maxdepth 1 -name "*.md" -mtime +90 -exec mv {} ~/meetings/archive/ \;") covering the common cases, with preview and apply variants shown. | 5 / 5 |
Workflow Clarity | The destructive workflow has a clear preview → user-confirmation → apply sequence ("Present this to the user before taking any action"; "Only apply cleanup after showing the preview and getting explicit confirmation"), so the destructive-cap doesn't apply, but there is no post-apply verification step or error-recovery loop to reach the top anchor. | 4 / 5 |
Progressive Disclosure | The single-file skill has well-organized sections (usage, cleanup tasks, gotchas) and no broken or buried references — no bundle files exist — but at ~90 lines it exceeds the simple-skill exception, and the Gotchas list is content that could be split into a reference file. | 4 / 5 |
Total | 18 / 20 Passed |