The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous command templates for offensive security tools, including the Impacket suite (psexec.py, smbexec.py, wmiexec.py, atexec.py, dcomexec.py), NetExec, and Rubeus, which are used to perform lateral movement and execute commands on remote Windows systems.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted engagement data from a state management server to inform its logic.
Ingestion points: The get_state_summary() function reads current engagement state from an MCP server.
Boundary markers: No delimiters or explicit instructions are provided to the agent to ignore potentially malicious instructions embedded within the retrieved state.
Capability inventory: The skill possesses high-privilege capabilities, including remote shell access and system modification tools across multiple protocols (SMB, WMI, WinRM).
Sanitization: The skill does not perform any validation or sanitization on the data received from the state server before incorporating it into command generation logic.
[COMMAND_EXECUTION]: Instructs the agent to perform administrative actions such as using sudo for local clock synchronization (ntpdate, rdate) and modifying remote system registry keys to enable specific authentication modes.

pass-the-hash