What problem does it solve? Running individual security checks in isolation leaves gaps: one tool audits RLS policies, another reviews authorization, and infrastructure threats go unchecked. This Skill chains five audit stages into one enforced sequence so a full-stack security review actually completes, with every finding consolidated into a single severity-ranked report. ## Core Features & Use Cases - Sequential 5-Stage Audit: Runs security-checklist (RLS, subscription state, rate limits, budget caps), authz-designer (endpoint authorization, IDOR scans, privilege escalation), paid-api-guard (six-layer defense for paid external APIs), /cso comprehensive (infrastructure, secrets, supply chain, STRIDE threat modeling), and /codex challenge (adversarial review). - Consolidated SUMMARY Report: Merges all stage outputs into docs/security/<date>-SUMMARY.md with issues sorted into Critical, High, Medium, and Low buckets. - Fix-and-Verify Loop: Drives atomic commits per issue, re-verification until Critical/High findings are resolved, and regression tests under tests/security/. - Use Case: Before a production release, ask for a full security audit; the Skill confirms scope and time budget, executes all five stages, writes per-stage reports, and blocks shipping until Critical and High issues are fixed. ## Quick Start Ask the agent to run a comprehensive security audit before deployment and confirm the audit scope and time budget when prompted.
