threat-modeling

name	threat-modeling
description	Performs data flow threat analysis with structured mitigations. Use when assessing security threats for new features, architecture changes, or compliance reviews.
argument-hint	[optional: --quick]

Threat Modeling

Reference

• ID: S-RISK-01
• Category: Risk
• Inputs: system architecture, data flow diagrams, trust boundaries, asset inventory
• Outputs: threat model document → artifacts/risk/
• Used by: Ops & Risk Agent
• Tool scripts: ./tools/artifact/validate.sh

Purpose

Produces a structured threat model that identifies attack surfaces, enumerates threats using a systematic methodology (STRIDE), and maps each threat to concrete mitigations, ensuring security is considered proactively rather than reactively.

Procedure

Step 0: Stage-Aware Mode Selection

Read company.stage from company.config.yaml. If idea or mvp:

MVP Security Checklist (use instead of full STRIDE):

1. Auth provider (not roll-your-own)
2. HTTPS enforced
3. Input sanitization on all user-supplied fields
4. Payment webhook signature verification (if applicable)
5. No secrets in code (run ./tools/security/secrets-scan.sh)
6. Dependency scanning configured

Reserve full STRIDE analysis for growth/scale stages or features touching financial data/PII at scale. If user passes --quick, always use MVP mode regardless of stage.

1. Define the scope: which system, feature, or data flow is being modeled.
2. Draw or review the data flow diagram (DFD): external entities, processes, data stores, data flows. Use the excalidraw skill (S-ENG-17) to produce the DFD — Mermaid flowchart LR with trust boundary subgraphs inline, and optionally an Excalidraw file for richer documentation.
3. Identify trust boundaries: where data crosses between trust zones.
4. Enumerate threats using STRIDE (Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, Elevation of Privilege) for each element.
5. Rate each threat: likelihood (1-5) and impact (1-5), calculate risk score.
6. For each threat, define a mitigation: preventive control, detective control, or accepted risk.
7. Prioritize mitigations by risk score.
8. Document residual risks: threats accepted without full mitigation.
9. Define review triggers: when this threat model should be revisited.
10. Save the threat model to artifacts/risk/.
11. Validate the artifact using ./tools/artifact/validate.sh.

Cross-References

• security-posture (S-RISK-04) — aggregate view of all threat model findings across features; run /security-posture for cross-cutting risk inventory
• release-readiness-gate (S-QA-03) — consumes threat model as a required artifact for Bar 3 (Security & Risk)

Quality Checklist

• Scope is clearly defined and bounded
• Data flow diagram is complete with all trust boundaries
• STRIDE analysis covers all DFD elements
• Each threat has likelihood and impact ratings
• Every high-risk threat has a defined mitigation
• Residual risks are explicitly documented and accepted
• Review triggers are specified
• Artifact passes validation