The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/model_usage.py invokes the codexbar CLI using subprocess.check_output(). The implementation is secure as it uses a fixed argument list and validates user-supplied input against a strict list of allowed options, preventing command injection.
[EXTERNAL_DOWNLOADS]: The skill's metadata includes installation instructions for the codexbar tool via a public Homebrew tap. This is an expected external dependency required for the skill's primary functionality.
[SAFE]: No malicious patterns such as data exfiltration, hardcoded credentials, or prompt injection were detected. The skill's operations are confined to processing local usage logs as described in its documentation.

model-usage