security-threat-model

Provide the repository path and any known context such as deployment model, internet exposure, and data sensitivity. The Skill maps components, trust boundaries, and entry points from repo evidence, then enumerates prioritized abuse paths and writes a Markdown threat model file.

It includes a system model with a Mermaid diagram, an assets table, an attacker model, entry points with evidence anchors, top abuse paths, a threat table with likelihood and impact ratings, mitigations, and focus paths for manual security review.

Use threat modeling when you need to understand attacker goals, trust boundaries, and prioritized abuse paths across the whole system. Use code review for line-level defect detection; this Skill explicitly does not trigger for general code review or architecture summaries.

No, but context improves accuracy. The Skill infers missing details, marks them as explicit assumptions, and asks 1-3 targeted questions about exposure, auth, and data sensitivity before producing the final report.

Findings depend on evidence discoverable in the repository, so undocumented infrastructure, runtime configuration, and organizational controls may be missed. Conclusions based on unresolved assumptions are marked conditional and should be validated by the service owner.