Conciseness

Roughly 33 lines, every sentence is dense instruction ('Answer these from the checkout', 'A skill that was never executed is a draft') with zero padding and no explanation of concepts Claude already knows. Clearly matches the lean, every-token-earns-its-place anchor.

Actionability

The core command is fully executable and verified real ('node <skill-dir>/scripts/write-verification-skill.mjs --code-root <code-root> --factory-root <factory-root> [--repo <repo-name>]'), the Drive section enumerates concrete surf commands, and FACTORY.json commands are specified at field level (name, argv, optional timeoutMs). Not a 5 because there is no copy-paste example of a completed FACTORY.json command entry and the semantics of --repo are only implied.

Workflow Clarity

Clear sequence — interview the repo, generate files, replace placeholders with observed values, run once, confirm evidence survives cleanup — with an explicit validation loop ('Fix the skill if that fails. A skill that was never executed is a draft.') and an explicit checkpoint ('confirm the evidence file is still there'). Matches the anchor with explicit validation and error-recovery feedback.

Progressive Disclosure

Under 50 lines with no need for external references: well-organized sections (## Interview, ## Prove one feature) and its single bundle reference (scripts/write-verification-skill.mjs) is real and correctly invoked, with the generated-skill detail appropriately delegated to that script. Per the rubric's simple-skill guidance this earns a 5.

Specificity

Names several concrete actions — 'Generate a project-local verification skill', 'drives the real app', 'captures evidence for factory lifecycle gates' — with minor gaps (the prove/fix loop and FACTORY.json check storage are unmentioned). Not a 5 because coverage of the skill's actions is incomplete; not a 3 because it lists well more than 1-2 concrete actions in a specific domain.

Completeness

Explicitly answers both: what ('Generate a project-local verification skill that drives the real app and captures evidence') and when ('Use when a project has no scripted way to prove UI, CLI, or service behavior') with a concrete trigger phrase. Clearly matches the anchor-5 example pattern of action list followed by an explicit 'Use when...' clause.

Trigger Term Quality

'prove UI, CLI, or service behavior' plus 'verification' and 'captures evidence' are terms a user would plausibly say ('no way to verify/test the UI'). Not a 5 because common synonyms like 'test', 'verify', 'e2e', or 'smoke test' are absent and 'factory lifecycle gates' is internal jargon; not a 3 because coverage goes beyond a single generic keyword.

Distinctiveness Conflict Risk

Clear niche (generating project-local verification skills for factory lifecycle gates) with a specific trigger unlikely to fire for unrelated skills. Not a 5 because 'prove UI, CLI, or service behavior' has minor overlap risk with general app-testing/run skills; not a 3 because the domain is far more specific than 'works with document files'.