agency-penetration-tester

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill document contains explicit offensive capabilities and step-by-step instructions (recon scripts, SQLi payloads, Kerberoasting/DCSync, NTLM/credential capture, reverse shells/tunneling, Meterpreter/Chisel/Ligolo usage, Golden/Skeleton/Shadow persistence techniques and exfiltration methods) that directly enable data exfiltration, credential theft, remote code execution, persistence/backdoors and network compromise, and thus poses a high malicious misuse risk.

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 includes explicit commands that modify the local host state and require elevated privileges (e.g., "sudo ip route add ...") and instructs starting local tunneling/C2 services, so it pushes the agent to change/privilege-escalate the machine it runs on.