performing-privilege-escalation-assessment

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill content and included script explicitly enumerate and instruct on multiple privilege-escalation and post‑compromise techniques — including commands to add root users, deploy reverse shells/MSI payloads, replace service binaries, exploit kernel/local vulnerabilities, and create persistent backdoors — which are dual‑use but clearly provide step‑by‑step actions that enable unauthorized system compromise and persistence if used without authorization.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The prompt explicitly instructs creating new root users, modifying system and service binaries/configuration, deploying kernel exploits, and other actions that directly modify or compromise the host system state, so it must be flagged.