threat_modeling

Threat modeling starts by identifying assets worth protecting, mapping trust boundaries where data crosses trust levels, then walking through STRIDE categories for each. This Skill automates that process and writes the result to threat_model.md in your project root.

STRIDE is a threat classification framework covering Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, and Elevation of Privilege. It provides a systematic checklist for reasoning about how each asset and trust boundary could be attacked.

Run it at project kickoff, before launches, after architectural changes like adding a database or external API, or when a feature introduces authentication or data storage. Skip it for cosmetic UI changes or refactors that do not alter trust boundaries.

It identifies likely threat patterns such as missing authorization checks, injection risks, and hardcoded secrets by reading the codebase, but it produces a reference document rather than executing exploits. It records required guarantees that guide later security work.

It should contain a project overview, an asset inventory, trust boundary descriptions, brief scan anchors for future reviews, and per-category threat narratives with required guarantees. Irrelevant STRIDE categories should be omitted rather than padded with boilerplate.