credential-dumping

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs extraction and reporting of plaintext credentials/tokens (e.g., printing "Password: $plaintext", saving credential dumps to evidence, and requiring the return summary to list "New credentials or tokens found"), which forces the agent to handle and output secret values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document is a deliberate, step‑by‑step playbook for credential theft and exfiltration against Active Directory (DCSync, NTDS/SAM/LSA dumps, ADSync DPAPI decryption, LAPS/gMSA abuses, persistence via DSRM/dMSA/GoldenGMSA, and remote execution/upload), i.e., clear malicious intent and high-risk abuse techniques.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs privileged, state-changing operations (uploading/executing scripts, creating VSS shadow copies and copying NTDS.dit, modifying registry and AD objects, creating scheduled tasks, changing ACLs/policies, etc.), which require elevated privileges and directly modify the host or domain state—so it pushes the agent to compromise the machine.