The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module across multiple utility scripts to execute system commands and external tools.
scripts/run_eval.py and scripts/run_loop.py execute the claude CLI to test generated skills and optimize descriptions.
eval-viewer/generate_review.py executes lsof and kill to manage the local web server port for the evaluation viewer.
scripts/package_skill.py executes python to run validation scripts before bundling skills into .skill files.
[EXTERNAL_DOWNLOADS]: The eval-viewer/viewer.html file includes a script reference to cdn.sheetjs.com to load the SheetJS library, which is used to render Excel (.xlsx) files within the evaluation report. This is a well-known service for client-side spreadsheet processing.
[DATA_EXPOSURE]: The eval-viewer/generate_review.py script reads the contents of output files from test runs and embeds them as Base64-encoded strings into a self-contained HTML file. This behavior is necessary for the skill's primary function of facilitating human review of generated content.

skill-creator