What problem does it solve? Passing happy-path tests does not mean a system is ready for real users or public exposure. This Skill performs an adversarial audit that finds real bugs, proves them with reproductions, fixes them, and leaves durable evidence before launch. ## Core Features & Use Cases - Seven-Surface Hunt Map: Systematically probes auth and sessions, authorization boundaries, input and upload handling, critical product journeys, API/MCP/agent surfaces, persistence and deployment assumptions, and visual integrity. - Proof Loop: Every suspected bug must be reproduced, covered by a failing test where practical, patched at the owning surface, and re-verified before being logged. - Safety Guardrails: Prohibits destructive probes against live data, requires confirmation before touching shared mounts, and mandates cleanup of started sessions. - Use Case: Before exposing a self-hosted app through a reverse proxy, run a full hunt to catch open redirects, cross-user data access, silent production fallbacks, and broken mobile layouts. ## Quick Start Run an adversarial bug hunt across this app before we expose it publicly, and prove, fix, and log every defect you find.
