Content
86%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.
The body is an exemplary lean instruction file: a concrete 4-step workflow, runnable commands, per-principle review questions, and a copy-ready report template with no filler. The only improvements are small robustness details — resolving the <plugin> path and changed-file-list placeholder and adding tool-failure handling.
Suggestions
Resolve the placeholders in the tool commands: state how to obtain the changed-file list (e.g., 'git diff --staged --name-only') and how the '<plugin>' path is determined, so the commands in step 2 are copy-paste executable.
Add a brief fallback for step 2 covering tool failure or empty output (e.g., what to do if complexity_checker or diff_surgeon errors out or returns no findings), giving the workflow a complete feedback loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence: 'You are opinionated and specific — don't just say "looks fine", point to exact lines and explain which principle they violate.' No concept explanations, no padding; every section (Role, Workflow, Rules) carries actionable weight, matching the 'every token earns its place' anchor. | 5 / 5 |
Actionability | Mostly executable guidance: concrete commands ('git diff --staged', 'python <plugin>/scripts/complexity_checker.py <changed-files> --json') and a complete report template with concrete review questions per principle. Minor gaps keep it below anchor 5: the '<plugin>' path placeholder is never resolved, and the workflow doesn't show how to obtain the changed-file list (e.g., 'git diff --staged --name-only'). | 4 / 5 |
Workflow Clarity | A clear numbered 4-step sequence (get diff → run tools → manual review → report) with one explicit fallback checkpoint ('If nothing staged, use git diff HEAD~1..HEAD (last commit)'). Below anchor 5 because there is no error-recovery guidance for tool failures or empty/unexpected tool output; this is a read-only review skill, so the destructive-operation validation cap does not apply. | 4 / 5 |
Progressive Disclosure | A single-purpose skill with well-organized sections (Role, a 4-step Workflow, Rules) where all content appropriately lives in SKILL.md — nothing needs splitting into reference files and there are no nested or buried references. Easy to navigate via clear headers, matching the well-organized-structure standard for simple skills. | 5 / 5 |
Total | 18 / 20 Passed |