ad-domain-attack

Warn

Audited by Gen Agent Trust Hub on May 9, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONCREDENTIALS_UNSAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides extensive command templates for tools such as Impacket, NetExec, BloodHound, and Kerbrute to perform domain enumeration and lateral movement.
  • [CREDENTIALS_UNSAFE]: Instructions are included for accessing and dumping sensitive files and memory regions, specifically targeting SAM hives (sam.hiv), SYSTEM hives (sys.hiv), and LSASS memory dumps (lsass.dmp) for credential extraction.
  • [DATA_EXFILTRATION]: Provides commands for extracting Active Directory secrets, including the NTDS.dit database and domain-wide NTLM hashes via DCSync, which are used for creating persistent access tokens like Golden Tickets.
  • [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface where the agent processes untrusted data. 1. Ingestion points: Results from network enumeration tools like bloodhound-python and ldeep (SKILL.md). 2. Boundary markers: None present in the instructions. 3. Capability inventory: Extensive command execution capabilities including impacket and bloodyAD scripts (references/domain-escalation.md). 4. Sanitization: No evidence of output validation or filtering of external tool results.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 9, 2026, 08:28 AM