offensive-security-analyst

SUSPICIOUS. The skill is internally consistent and not malware-like, but it gives an AI agent offensive security capabilities with real-world impact. Main risk is misuse of authorized pentest guidance, not credential theft or exfiltration.