Content
100%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 SKILL.md body for a simple command-reference skill: terse, fully executable examples, sensible sections, and no padding or concepts Claude already knows. Nothing material to improve within its scope.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean — every section is a command pattern with a one-line purpose ('Symbol definition: rg -n "fn\s+\w+\(" src/'), with no explanation of what regex or git is. The only explanatory line ('The built-in grep tool matches LITERAL text only') is environment-specific and earns its place. Nothing can be trimmed; this matches the 'every token earns its place' anchor. | 5 / 5 |
Actionability | All guidance is copy-paste-ready: 'rg -n --type java "class \w+Service"', 'git log -S "<string>" --oneline', 'git blame -L <start>,<end> <file>', plus concrete flags (-F, -l, --files). Placeholders like <start>,<end> are legitimate parameters, and the examples cover the common cases (definitions, usages, file-type filtering, file listing, history). | 5 / 5 |
Workflow Clarity | This is a simple, single-purpose reference skill (~29 lines) with unambiguous per-task commands, so the simple-skill exception applies and workflow clarity can score 5. It involves only read-only searches, so no validation/feedback-loop cap applies; sections sequence cleanly from regex search to history search to tips. | 5 / 5 |
Progressive Disclosure | The skill is under 50 lines with no external references needed, and no bundle files (references/scripts/assets) exist, so none are required. The body is well-organized into clear sections (Ripgrep, Git history, Tips) — matching the guideline that short, self-contained skills score 5 on well-organized sections alone. | 5 / 5 |
Total | 20 / 20 Passed |