| name | threatModel |
| description | Use when: producing STRIDE threat models with trust boundaries, asset identification, and feature threat classification. |
| type | reference |
| version | 1.0 |
| license | MIT |
Threat Model
Skill metadata: version "1.0"; tags [security, threat-modeling, stride]; recommended tools [].
Use this skill in workspaces with the secure pack selected.
Produce lightweight STRIDE threat models for features and modules. A threat model identifies what can go wrong before writing code, not after.
When to use
- Modeling a new user-facing feature, API endpoint, authentication flow, or integration with external services
- Any feature handling sensitive data, authentication, authorization, or with financial or compliance implications
When NOT to use
- When reviewing already-deployed, stable, unchanged infrastructure
- When reviewing code style or documentation only
When to threat model
- A new user-facing feature or API endpoint
- Any feature handling authentication, authorization, or sensitive data
- Integrations with external services or data stores
- Features with financial, legal, or compliance implications
STRIDE categories
| Threat | Description | Example |
|---|
| Spoofing | Pretending to be someone else | Forged auth tokens, impersonation |
| Tampering | Modifying data without authorization | Input manipulation, log tampering |
| Repudiation | Denying an action occurred | Missing audit logs |
| Information Disclosure | Exposing data to unauthorized parties | Verbose errors, IDOR |
| Denial of Service | Exhausting resources | Missing rate limits, unbounded input |
| Elevation of Privilege | Gaining unintended permissions | IDOR leading to admin access, SSRF |
Threat model output format
Produce this table for every threat model:
| Threat type | Description | Likelihood (H/M/L) | Mitigation |
|---|
Followed by:
- Trust boundaries identified: list the points where data crosses a trust boundary (e.g., user input → backend, backend → database, service → external API)
- Assets at risk: what data or capabilities an attacker would target
Scope discipline
Keep the model focused on the component described. Do not enumerate every theoretical risk in the broader system. One table per component is correct; cross-component threats belong in a separate model for the integration point.
Likelihood guidance
- High: straightforward to exploit with widely available tools or knowledge
- Medium: requires meaningful effort or specific knowledge
- Low: requires unusual conditions, insider access, or sophisticated capability
Verify