ad-delegation-attack
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides comprehensive instructions for executing security tools via the command line across Linux and Windows platforms. This includes using Impacket scripts (getST.py, rbcd.py), Rubeus, Coercer, and PetitPotam for tasks such as ticket manipulation, Active Directory object modification, and triggering forced authentication. These operations are fundamental to the skill's stated purpose of security auditing.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface where the agent processes external data sources which could contain malicious instructions.
- Ingestion points: The agent is instructed to interpret results from BloodHound and output from tools like findDelegation.py, crackmapexec, and rbcd.py (SKILL.md, references/rbcd-attack.md).
- Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to treat data from the target Active Directory environment as untrusted.
- Capability inventory: The skill requires significant system capabilities, including command execution and network interaction with domain controllers, to perform the described tasks.
- Sanitization: The skill does not provide methods or instructions for sanitizing or validating tool outputs before they are processed by the agent to determine subsequent actions.
Audit Metadata