Content
93%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: dense with executable tool calls and commands, tight prose, and explicit verification and error-recovery checkpoints. The only soft spot is that its workflow is presented as situation-indexed guidance rather than a single ordered sequence.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — it lists markers ("McCabe complexity, deep nesting, brain methods, class cohesion (LCOM4), god classes, clone detection") without explaining them and never teaches background concepts; only trivially trimmable redundancy remains ("Zero LLM calls; pure local analysis" vs "no LLM needed"). | 5 / 5 |
Actionability | Fully executable, copy-paste-ready guidance with literal arguments: get_health(targets=["src/x.py", "src/y.py"]), get_health(fix_id=...), repowise coverage add cov.lcov, repowise health --recompute, and /repowise:init, plus exact include-flag values. | 5 / 5 |
Workflow Clarity | Mode selection is clearly disambiguated and the results-usage list includes genuine checkpoints ("Check unresolved before calling a file clean", "Before editing a flagged file → cross-check get_risk") and error-recovery paths (not_indexed → repowise update; no repository → /repowise:init), but the guidance is organized as parallel situational playbooks rather than one strictly sequenced workflow. | 4 / 5 |
Progressive Disclosure | No bundle files exist, and the ~58-line body is appropriately self-contained: it inlines nothing that belongs in a separate file and is organized into clear, navigable sections (modes, include flags, results usage, CLI equivalents, error handling), fitting the simple-skill exception. | 5 / 5 |
Total | 19 / 20 Passed |