The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts agent.py and process.py automate calls to external security tools (such as NetExec, Impacket, and BloodHound) using subprocess.run(). This is the primary purpose of the skill and facilitates Active Directory enumeration and exploitation tasks.\n- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface in process.py, where output from the netexec tool is parsed to generate a list of domain users, which is then used as input for subsequent tool executions.\n
Ingestion points: The enumerate_domain_users function in scripts/process.py parses stdout from the netexec command.\n
Boundary markers: None present to delimit tool output from agent instructions.\n
Capability inventory: Execution of various security CLI tools via subprocess.run, file system writes for user lists and reporting, and LDAP queries via ldap3.\n
Sanitization: Basic string splitting and filtering is performed, but no robust validation of the parsed usernames is implemented before their use in subsequent file-based tool inputs.

performing-active-directory-penetration-test