The COMPREHENSIVE tier. Build (or ingest) a full STRIDE threat model as input, then run a sophisticated whole-repo adversarial zeroday exploit review (triple-pass HUNT -> JUDGE -> VERIFY) that validates the model against the code and proves a runnable PoC per confirmed finding. Composes the threat-modeling + 0day-scanner (deep) skills. Use for "tm security review", a deep exploit-focused review with a threat model, or any agentic/LLM/MCP codebase where design risk outweighs pattern bugs. Ask about deployment/trust context when unclear. For a quick pre-merge diff pass instead, use tm-security-assessment.
