Windows Privilege Escalation

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs harvesting and embedding cleartext credentials (e.g., "DefaultPassword: P@ssw0rd123" and commands using -p P@ssw0rd123 or netsh ... key=clear), which requires the LLM to handle and output secret values verbatim, creating an exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document contains explicit, actionable instructions for credential harvesting, remote command execution (reverse shells), privilege escalation exploits, and AV-evasion—demonstrating clear malicious intent and guidance for backdoor-style access.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). This skill explicitly instructs modifying system services, registry, and system files and running privilege-escalation exploits (sc config, msiexec, JuicyPotato, mimikatz, SAM/SYSTEM extraction, WSL root setting), which directly compromise and change the machine's state.