The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides an extensive collection of command-line templates for executing Impacket scripts, including psexec.py, wmiexec.py, ntlmrelayx.py, and secretsdump.py. These allow the agent to perform remote command execution and interact with various Windows protocols.
[EXTERNAL_DOWNLOADS]: The instructions reference the official Impacket repository on GitHub (github.com/fortra/impacket) for installation and also suggest installing the coercer tool via pip. These are well-known resources in the security community.
[DATA_EXFILTRATION]: The skill documents methods for extracting highly sensitive data, such as dumping NTLM hashes from the NTDS.dit database, SAM hives, and LSA secrets. This behavior is the intended primary function of the toolkit for Active Directory auditing.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits a vulnerability surface where untrusted data could be interpolated into shell commands via placeholders.
Ingestion points: Commands use placeholders for user-supplied data such as DOMAIN/user:password@TARGET and payload strings in -c "PAYLOAD" arguments.
Boundary markers: There are no explicit markers or instructions to the agent to sanitize or ignore embedded instructions within data used to fill these placeholders.
Capability inventory: The skill provides a high-privilege capability set including remote system access, credential harvesting, and network traffic relaying.
Sanitization: No input validation or escaping logic is provided for the command-line parameters.

impacket