ad-domain-attack
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill documents the use of numerous security assessment tools (e.g., netexec, impacket, bloodhound-python, certipy) which involve executing shell commands to interact with network services and target systems.
- [INDIRECT_PROMPT_INJECTION]: The skill interacts with external environment data through various enumeration tools. Maliciously crafted data in the target environment (e.g., special characters in Active Directory attributes or file names) could potentially influence the agent's behavior if the tool outputs are processed without sanitization.
- Ingestion points: Output from network enumeration tools like netexec, bloodhound-python, and ldeep (found in SKILL.md and references/authenticated-enum.md).
- Boundary markers: No specific delimiters are suggested for handling external tool output.
- Capability inventory: Extensive command execution capabilities including remote shell access (impacket-psexec, evil-winrm) and domain modifications (bloodyAD).
- Sanitization: The skill does not provide methods for sanitizing data retrieved from the target environment.
- [CREDENTIALS_UNSAFE]: The documentation includes common password lists for spraying and a well-known, publicly available AES key used to decrypt legacy Group Policy Preferences passwords. These are provided as part of the penetration testing methodology.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the standard krb5-user package from the official system repository to support Kerberos authentication.
Audit Metadata