The Agent Skills Directory

[COMMAND_EXECUTION]: The skill explicitly instructs the agent to run scripts/invoke.py and adopt the output as standing instructions for the session, which is the primary mechanism for switching to delegation mode.
[DYNAMIC_EXECUTION]: The scripts/invoke.py wrapper dynamically searches for a central management engine (kntnt.py) by iterating through a pre-defined list of standard AI agent skill directories in the user's home folder, executing it via subprocess.run.
[INDIRECT_PROMPT_INJECTION]: The skill takes verbatim user input following the /delegation command and passes it to the external management engine via stdin, creating a risk that malicious input could influence the instructions returned to the agent. 1. Ingestion points: The natural language instructions following the command separator in SKILL.md. 2. Boundary markers: No delimiters or ignore instructions are applied to the user input before processing. 3. Capability inventory: Use of subprocess.run in scripts/invoke.py and instructions in SKILL.md for the agent to follow the resulting output. 4. Sanitization: The provided implementation does not include validation or sanitization of the passed user input.

delegation