Conciseness

The body is information-dense (per-mutator separating inputs, tool quirks, score conventions) with little padding, but at ~300 lines it has trimmable sections: the Step 4 example block embeds a discussion paragraph ('If a boundary test already exists...') inside the template, and the Step 5 prose could be tightened. Between anchors 4 and 5 — efficient with minor instances of over-explanation.

Actionability

Fully executable guidance: copy-paste-ready TypeScript for the SurvivedMutant record, four worked mutation families each with concrete separating inputs and proposed fixes, and a complete worked example test (`expect(() => cart.addItem({ qty: 100 })).not.toThrow()`) plus an exact output-format template. This is an instruction-only skill and the guidance is fully actionable.

Workflow Clarity

Clear five-step sequence (normalize → classify → heuristics → propose → equivalence) with an explicit rationale for step ordering, a four-item completeness checklist in Step 4 acting as validation, and error-recovery branching ('If a boundary test already exists and the mutant still survived, either the assertion does not observe the throw... or the production boundary is off by one'). The skill is explicitly read-only ('Never auto-rewrite tests'), so the destructive-operation cap does not apply.

Progressive Disclosure

One reference file (references/tool-normalization.md), clearly signaled with two inline links and a summary of what it carries, exactly one level deep and verified to exist with substantive per-tool tables. The body keeps the workflow and offloads per-tool field mappings and operator-name tables appropriately.

Specificity

Multiple specific concrete actions with comprehensive coverage: 'Normalizes a surviving-mutant record... classifies why it survived (missing case, weak assertion, equivalent mutant, unreachable code, flaky killer), applies per-mutator heuristics for conditional-boundary, arithmetic-operator, statement-removal, and constant mutations, and drafts the specific test that would kill it.' Matches the 5 anchor; the enumerated classes and mutator families leave no minor gaps that would drop it to 4.

Completeness

Explicitly answers both: the what ('normalizes... classifies... drafts the specific test') and the when ('Use when a mutation run has finished and the report lists surviving mutants (typically 5+) that nobody has yet explained or turned into concrete test cases') with a concrete trigger condition.

Trigger Term Quality

Comprehensive natural-term coverage: 'surviving mutants', 'mutation run', 'mutation report', tool names (StrykerJS, PIT, mutmut, Mull), and report formats ('Stryker JSON / PIT XML / mutmut output / Mull JSON'). These are the exact phrases a user with a finished mutation run would say, including tool-name synonyms.

Distinctiveness Conflict Risk

Clear niche (post-run surviving-mutant triage) with distinct triggers — no other plausible skill fires on 'mutation run has finished and the report lists surviving mutants'. Also correctly scopes itself against neighbors ('never auto-rewriting tests', starts at the report, ends at a proposal), further reducing misfire risk. Third-person voice throughout.