pentest-engagement-methodology

SUSPICIOUS: this is a methodology-only skill with no obvious malware, hidden execution, or credential-harvesting path, but it materially equips an AI agent to perform offensive security actions end-to-end. Its purpose and content are internally consistent, yet the capability itself is high risk for an agent deployment.