performing-security-audits

This skill’s stated purpose broadly matches its capabilities, but it is a high-risk security/audit skill because it grants unrestricted shell access and references opaque plugin subtools without clear execution or data-flow boundaries. I see no confirmed malicious behavior, but the combination of AI-driven security auditing, broad permissions, and unspecified backend/tool behavior makes it suspicious rather than benign.