What problem does it solve? Deciding whether a finished QA or verification test plan is good enough to execute is subjective and error-prone: reviewers rubber-stamp long catalogs, miss behaviors with no test case, or over-flag sound plans. This Skill acts as an independent acceptance gate that judges a test plan against a fixed 10-condition bar and returns a machine-parseable verdict with actionable findings. ## Core Features & Use Cases - Coverage completeness check: Enumerates every behavior, operation, and error case from the handed-in upstreams (feature-spec, api-spec, or PRD) and confirms each has at least one traceable test case. - Risk-weighted catalog judgment: Detects coverage gaps, combinatorial blow-ups, and padded or thin catalogs, plus checks entry/exit criteria, environments, test data, and warranted non-functional targets such as WCAG 2.2 AA and p95 latency. - Delta-scoped amend review: When a change request is handed in, judges only the delta, coverage re-tracing, and regression selection rather than re-reviewing the whole plan. - Use Case: After authoring a test plan for a checkout feature, run this gate with the plan plus the feature-spec and api-spec; it returns VERDICT: revise naming the uncovered behaviors and missing performance targets, or VERDICT: approve if every applicable condition passes. ## Quick Start Use reviewing-test-plan on docs/qa/checkout-test-plan.md against the feature-spec and api-spec, and give me the verdict with findings.
