Security playbook for building, auditing, and repairing AI-generated or "vibe coded" systems. Use when reviewing vibe-coded apps, writing security-first prompts, hardening Supabase/RLS/API flows, designing pentest checklists, or repairing vulnerabilities such as race conditions, IDOR, user enumeration, upload validation flaws, secrets exposure, and business-logic abuse.
This skill includes a SKILL.md with agent instructions.