What problem does it solve? Security prior-art surveys produce threat-vocabulary maps, per-angle search outputs, extract records, and threat registers whose honesty cannot be verified by schema validation alone. This Skill acts as an independent acceptance gate that judges whether the research craft behind those artifacts is honest, complete against its own contracts, and proportionate, so a dishonest coverage record does not become a confidently wrong claim about what threats exist. ## Core Features & Use Cases - Deterministic-first review: Runs the co-installed producer's validator once, then judges what a validator cannot see, such as whether a coverage claim is provable or a source failure was written as a zero. - Condition-based findings: Applies the numbered bar in references/conditions.md across all artifact kinds and emits VERDICT: approve or revise with condition-named findings and concrete locations. - Spot-check posture: Samples coverage cells and candidates to confirm recorded queries plausibly yield recorded results, without re-deriving the whole search. - Use Case: After authoring a threat register with the security-prior-art-survey producer, run this gate to verify coverage completeness, sanitization records, evidence-tier grounding, and vocabulary-based threat naming before approving the survey. ## Quick Start Use reviewing-security-prior-art-survey to judge the threat register in docs/security/register.md and report any blocking findings.
