pentesting

SUSPICIOUS: the content is internally consistent with a pentesting/reporting skill, but it grants an AI agent offensive security capability beyond passive documentation. No clear malware, exfiltration path, installer abuse, or credential harvesting is present; the main risk is enabling recon/exploitation workflows if the agent has broad execution permissions.