Conciseness

The body is lean throughout: it assumes Claude already knows TDD and adds only policy — 'If a test passes before implementation, it is not red evidence', 'Review a slowdown only when it exceeds both 20 percent and 100 ms'. There is no padding, no conceptual explanation of what tests or refactoring are, and every sentence constrains behavior.

Actionability

Guidance is highly concrete for an instruction-only skill: a numbered six-step rule, named ledger fields ('old_test', 'behavior', 'replacement', 'mutant', 'red_observed', 'baseline_ms', 'candidate_ms', 'reason'), and exact measurement protocol (one warm-up, five isolated runs, median). It stops short of a 5 because no example of the ledger or a red-evidence record is given, so the exact output format is left implicit.

Workflow Clarity

The sequence is explicit with validation checkpoints at each gate: run the focused test and record the expected failure and exit status, implement only after red, run the focused test then the affected suite, refactor only while green. Feedback loops are present — repairing a test that errors 'until it fails on the missing behavior', and stopping after two failed attempts to re-check the boundary — satisfying the rubric's error-recovery requirement.

Progressive Disclosure

The skill is a short, single file with well-organized sections (rule, consolidation, strategy rotation) and no bundle directory, so there is nothing that should be split out. It loses a point because the body references files that are not part of this skill's bundle — 'skills/blocks/engineering-method-selection.md' and 'THIRD_PARTY_NOTICES.md' — leaving two dangling paths a reader cannot resolve locally.

Specificity

The description lists several concrete actions — 'Build a behavior change', 'observed red, minimal green', 'measured test consolidation' — each naming a distinct step of the TDD cycle with modifiers that constrain how it is done. It falls short of a 5 because coverage is compressed into jargon shorthand and omits parts of the cycle (e.g., refactor), leaving minor gaps.

Completeness

The 'what' is clearly stated — build a behavior change through observed red, minimal green, and measured test consolidation — but the description itself contains no 'Use when...' clause or equivalent explicit trigger guidance, capping completeness at 3 per the rubric. (The frontmatter has a separate 'trigger:' field, but the description field evaluated here does not answer 'when' on its own.)

Trigger Term Quality

It contains some relevant keywords ('behavior change', 'red', 'green', 'test consolidation') but misses the natural phrases most users say, such as 'test-driven development', 'TDD', 'write tests first', or 'unit tests'. It sits above 2 (more than one or two generic keywords) but below 4, which requires good coverage of natural terms users would actually say.

Distinctiveness Conflict Risk

The red/green/consolidation language carves out a clear TDD niche that is unlikely to trigger for unrelated skills, but the opening phrase 'Build a behavior change' is broad and the description lacks the explicit 'test'/'TDD' naming that would fully separate it from general coding or testing skills, keeping it just below a 5.