active-directory-attacks

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides a large number of command-line examples for executing powerful security tools such as Mimikatz, Rubeus, Impacket, and BloodHound. These instructions allow for system-level manipulation and network-wide exploitation.
  • [DATA_EXFILTRATION]: Contains detailed procedures for extracting sensitive authentication material, including NTLM hashes, Kerberos tickets (Golden/Silver tickets), and service account passwords. While the instructions focus on local extraction to files like 'hashes.txt' or 'tgs.txt' for testing, they represent a significant exposure of sensitive domain credentials.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its interaction with external data.
  • Ingestion points: Active Directory object attributes (users, groups, GPOs) enumerated through tools like BloodHound, PowerView, and certipy (SKILL.md, references/advanced-attacks.md).
  • Boundary markers: No delimiters or safety instructions are used to prevent the agent from obeying malicious instructions that might be embedded in Active Directory data fields.
  • Capability inventory: Extensive system and network capabilities including subprocess execution of security tools, network exploitation (Kerberoasting, DCSync), and local file creation (SKILL.md).
  • Sanitization: No methods are described for sanitizing or validating data retrieved from the AD environment before it is used by the agent or presented to the user.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 08:59 AM