The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill includes a discovery script (invoke.py) that searches for a management engine named kntnt.py across a wide predefined list of local directory paths associated with various AI agent harnesses (e.g., Claude, Cursor, Aider). This mechanism allows the skill to integrate with its parent framework regardless of the specific environment where it is deployed.
[COMMAND_EXECUTION]: The skill invokes the discovered management engine via subprocess.run using the uv tool. The implementation passes parameters as an argument list and channels user input via standard input (stdin), which is a secure method that avoids shell injection vulnerabilities and correctly scopes the execution to the local system.

push