physical-security-testing
Audited by Socket on Sep 19, 2026
3 alerts found:
Securityx2MalwareThis skill is an offensive physical-intrusion playbook for an AI agent, with explicit instructions for badge cloning, covert entry, implant deployment, reverse shells, credential harvesting, and hidden surveillance. Its capabilities are coherent with its stated purpose, but that purpose itself gives the agent high-risk real-world attack functionality and exfiltration paths; classify as suspicious/high-risk vulnerable rather than confirmed malware.
The fragment is a readable physical-security penetration-testing playbook. It does not itself execute code or demonstrate npm supply-chain malware, but it documents actionable methods for credential cloning, covert network implants, reverse shells, surveillance, and social engineering. Claimed authorization and cleanup controls reduce contextual concern but do not eliminate the high misuse potential of the procedures.
The fragment contains explicit offensive malware-like capabilities and operational instructions for covert deployment. It establishes persistent reverse SSH access, executes arbitrary commands through hidden PowerShell reverse shells, supports DNS and SCP exfiltration, downloads and runs remote payloads, captures and uploads images, modifies system binaries, clones access credentials, and provides concealment and evasion guidance. Even if presented as authorized red-team material, these mechanisms are directly usable for unauthorized compromise and surveillance. The package or file should be treated as dangerous and should not be executed or distributed without strict isolation and verified authorization.