active-directory-attack

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt repeatedly instructs embedding plaintext credentials, passwords, and hashes directly into commands and tool parameters (e.g., user:password, -p 'Password123', --ldappassword, /credpassword, NTLM_HASH/HASH placeholders), which forces the agent to handle and potentially 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 content is an explicit offensive playbook detailing step‑by‑step techniques for credential theft (Kerberoasting, AS‑REP, DCSync), lateral movement and remote execution (Pass‑the‑Hash/Ticket, NTLM relay, psexec/wmiexec), and persistent backdoors and domain compromise (Golden/Silver/Diamond/Sapphire tickets, Skeleton Key, DSRM backdoor, DCShadow), i.e., clearly intended to enable malicious Active Directory compromise.

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 actions that alter system and domain state and require elevated privileges (e.g., "sudo date -s", Set-ItemProperty to change DSRM behavior, Add-DomainObjectAcl, account/ACL creation via relay and Mimikatz), so it pushes the agent to compromise the host/environment.