Conciseness

The body is lean with zero padding: every bullet is a dense detection condition ('a touched file pushed past 1000 lines', 'more than two delegation hops', 'booleans without is, has, or should') and no concept Claude already knows is explained. This matches anchor 5 (lean and efficient, every token earns its place).

Actionability

For an instruction-only skill, the guidance is fully actionable: concrete thresholds per smell, a concrete method ('read the sibling helpers and their comments', 'Read the AGENTS.md or CLAUDE.md chain governing the changed files'), and a specific output spec ('Quote the shape... Give a concrete reframe: what to delete, split, move, or collapse, not "consider refactoring"'). This matches anchor 5 for instruction skills with concrete, specific guidance.

Workflow Clarity

This is a simple, single-purpose skill (a review lens) whose flow is unambiguous across Scope, Method, Threshold, and Reporting sections, so the simple-skill exception applies and workflow clarity scores 5. It is a read-only review task, so the destructive/batch validation cap does not apply, and the Threshold section supplies an explicit decision checkpoint for what to report and what to suppress.

Progressive Disclosure

The body is under 50 lines, requires no external references, and is cleanly organized into four well-labeled sections (Scope, Method, Threshold, Reporting), which meets the guideline for simple skills to score 5 with just well-organized sections. No bundle files exist to reference, and none are needed.

Specificity

The description names one concrete action ('Review a code change for structural regressions that make code harder to change, delete, or reason about') and enumerates many specific finding types ('complexity moved rather than removed', 'shared-path special cases', 'oversized files', 'thin wrappers', 'premature abstraction', 'dead code'). It falls at anchor 4 rather than 5 because it lists specific review targets under a single verb rather than multiple distinct concrete actions like the anchor-5 example.

Completeness

Both 'what' (review code changes for the enumerated structural regressions) and 'when' (an explicit 'Use when reviewing for maintainability, simplicity, code structure, or design quality' clause with concrete trigger phrases) are clearly and explicitly stated, matching the anchor-5 pattern of the good overall examples. It is not anchor 4 because the 'when' is not merely present but specific and trigger-phrase-based.

Trigger Term Quality

The 'Use when reviewing for maintainability, simplicity, code structure, or design quality' clause supplies natural phrases users would say. It matches anchor 4 (good keyword coverage, a few natural terms missing) rather than anchor 5 because common synonyms like 'refactor', 'code smells', or 'clean code' are absent.

Distinctiveness Conflict Risk

The maintainability lens is a clear niche ('structural regressions that make code harder to change, delete, or reason about'), but the opening 'Review a code change' overlaps with general correctness-review skills and 'design quality' is a broad trigger, so it fits anchor 4 (mostly distinct, minor overlap risk with closely related skills) rather than anchor 5.