Use this skill to operate a hardened AI agent that analyzes vulnerabilities in authorized systems, challenges weak evidence, builds attack paths, prioritizes real risk, and recommends safe validation and remediation