What problem does it solve? Security review often stops at prose opinions, leaving no executable proof that abuse paths were actually exercised. This Skill turns a threat model into concrete, reproducible security tests, tool runs, and evidence records for a pinned code change. ## Core Features & Use Cases - Threat-model-driven test selection: Consume or create a minimal threat model covering assets, trust boundaries, attacker capabilities, and fail-closed expectations before choosing tests. - Executable abuse-case coverage: Run or add tests for authz bypass, input validation, injection, path traversal, env/command handling, unsafe deserialization, secrets/log hygiene, and dependency or supply-chain exposure. - Structured qa-result evidence: Emit a standardized evidence record with status, reviewed ref, scope, risk class, tests executed, and residual risk. - Use Case: After modifying a CLI that parses user-supplied file paths, use this Skill to build a minimal threat model, run hostile-input probes against local fixtures, and emit a qa-result record proving path traversal is blocked. ## Quick Start Ask the agent to run security testing on the current change, covering input validation and path handling, and emit a qa-result evidence record.
