The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the subprocess and os modules to automate developer workflows. Specifically:
scripts/run_eval.py uses subprocess.Popen to execute the claude CLI for trigger evaluation tests.
scripts/improve_description.py invokes claude -p to generate optimized skill descriptions.
eval-viewer/generate_review.py runs lsof and uses os.kill to manage local network ports for the evaluation viewer server.
These actions are standard for developer tooling and are performed within the local environment context.
[EXTERNAL_DOWNLOADS]: The eval-viewer/viewer.html template references the SheetJS (xlsx) library from cdn.sheetjs.com. This is a well-known and trusted external service used for processing spreadsheet data within the browser-based results viewer.
[SAFE]: The skill generates temporary markdown files and JSON metadata in .claude/commands/ and specified workspace directories to facilitate testing. This dynamic file creation is necessary for its function as a skill-testing harness and does not exhibit malicious patterns.

skill-creator