The Agent Skills Directory

[SAFE]: The skill implements security guardrails intended to monitor and restrict agent actions, enhancing the safety of the execution environment.- [COMMAND_EXECUTION]: The file scripts/test_hooks.py uses subprocess.run to call the skill's own Python scripts for verification during development. This execution is limited to internal scripts and uses controlled inputs for automated testing.- [DYNAMIC_EXECUTION]: The scripts/guard_one_way_doors.py script opens and reads local script files referenced in shell commands to check for restricted API calls. This is a read-only operation performed as a necessary security check for the hook's protective function.

guardrail-hooks