Conciseness

The body is lean, lists marker names ("McCabe complexity, deep nesting, brain methods, class cohesion (LCOM4), god classes") without explaining what they are, and earns its one explanatory line ("a low score means more likely to harbour bugs, not just bigger") by correcting a likely misreading. It is not a 5 because the include-flags section and the CLI-equivalents section overlap somewhat and could be tightened.

Actionability

Concrete, executable guidance throughout: tool-call signatures "get_health()", "get_health(targets=[\"src/x.py\", \"src/y.py\"])", exact include/only flag strings, and copy-paste CLI commands ("repowise health --refactoring-targets", "repowise coverage add cov.lcov") covering both dashboard and targeted modes. This matches the anchor for fully executable, common-case-covering guidance.

Workflow Clarity

"How to use the results" is a clear six-step decision guide with explicit checkpoints ("Check unresolved before calling a file clean", "not_indexed means run repowise update", "cross-check get_risk") and a feedback loop (ingest coverage then re-run repowise health). It is a usage guide rather than a strict linear pipeline, and a couple of checkpoints are implicit, so it sits at 4 rather than 5.

Progressive Disclosure

A single-purpose skill (~57 lines) with no need for external bundle files, organized into clearly headed sections (Pick the mode, Useful include flags, How to use the results, CLI equivalents, Error handling) and no nested references. Per the simple-skill guidance, well-organized sections with no external-reference need warrant a 5.

Specificity

The description names the domain richly ("code health, code quality, complexity, technical debt, ... untested hotspots, or coverage gaps") but enumerates user concerns rather than the skill's concrete actions; the only action-ish phrase is "get a before/after health read". It sits above anchor 2 (which would be a single generic action) because the domain topics are specific, but below 4 because the skill's actual capabilities (scoring files, returning findings, ranked refactoring plans) are not stated.

Completeness

The "when" is exceptionally explicit ("Use when the user asks about... Also use to get a before/after health read when planning or finishing a refactor"), but the "what" is only weakly implied via "code health" / "health read" rather than a concrete capability statement. Both are present so it clears anchor 4, but it is not a 5 because the what is not explicitly and concretely stated.

Trigger Term Quality

Phrases like "code health", "code quality", "complexity", "technical debt", "which files are risky or hard to maintain", "what to refactor next", "untested hotspots", "coverage gaps", and "before/after health read" are exactly what a user would naturally say, with strong synonym coverage. This matches the anchor for comprehensive natural-term coverage including synonyms.

Distinctiveness Conflict Risk

Scoping to a "Repowise-indexed codebase (.repowise/ directory exists)" gives it a clear niche and low conflict risk, but generic triggers like "what to refactor next" and "code quality" could overlap with general refactoring/code-quality skills, so it is mostly distinct with minor overlap rather than fully conflict-free.