network-recon

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill contains explicit offensive techniques (LLMNR/NBNS poisoning via responder), instructions to obtain and use remote shells/pivot hosts, and step-by-step methods to upload and run binaries on pivot hosts (base64/certutil/python HTTP transfer, running static nmap), plus proxychains-based evasion — all high-risk dual-use actions that enable remote code execution, internal network compromise, and covert operations if abused.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill explicitly instructs disabling the execution sandbox ("dangerouslyDisableSandbox: true") and includes repeated use of sudo nmap handoffs and uploading binaries to pivot hosts, which directs bypassing security controls and modifying system state.