security-architecturelisted

# Security Architecture Owns the system's secure DESIGN: trust boundaries, identity & access model, zero-trust enforcement points, multi-tenant isolation, data-protection/key strategy, and the threat model. Honors `${CLAUDE_PLUGIN_ROOT}/skills/_shared/vibegod-principles.md`. Priority order: **user > skills > default behavior** — never silently override the user. You set the secure baseline the build must follow; you do not perform code-level review. ## Fits in the pipeline Serves the **Architecture department (Stage 4 — module map & blueprint)** and feeds the **Stage 5 build plan** with security acceptance criteria. Threat modeling is continuous, not a one-time gate: any new service, trust boundary, tenancy model, or auth change re-enters here. The implementation-time OWASP lens lives in the **Stage 7 per-feature QA gate** (security-engineer + `secure-coding`) — this skill writes the design that lens checks against. ## Best practices **Zero Trust (NIST SP 800-207 §2.1 tenets):** - Treat every data source and service as a **resource to protect** — nothing exposed without an explicit policy. - **Secure all communication regardless of network location** — "internal network" / VPN never grants implicit trust. - Grant access **per-session, per-request, least privilege**, re-evaluated each time — never "authenticate once, trust forever". - Make access decisions **dynamic** — from client identity, service identity, device posture, and observable attributes/behavior. - Complete *