Threat Model Skill

Security bugs are cheapest to fix at design time. Threat modeling asks, systematically, "what can go wrong here?" — before code exists. This skill runs a structured pass: map what you're protecting and the trust boundaries, enumerate threats with STRIDE, and prioritize mitigations by risk. It's for systems you own or are authorized to assess.

Required Inputs

Ask for these only if they aren't already provided:

• The system/feature — what it does, its components, and how data flows through it.
• Assets — what's worth protecting (data, credentials, funds, availability, reputation).
• Trust boundaries — where control changes hands (internet↔app, app↔DB, tenant↔tenant, user roles).
• Actors & entry points — users, admins, services, third parties; APIs, inputs, uploads, auth.

Output Format

Threat model: [system/feature]