The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Python's subprocess module to interact with the host system and the platform's CLI tool.
scripts/run_eval.py and scripts/improve_description.py execute the claude CLI to run evaluations and generate improved skill descriptions based on test results.
eval-viewer/generate_review.py executes the lsof command to identify and manage network ports used by the local viewer server.
[EXTERNAL_DOWNLOADS]: The evaluation viewer (eval-viewer/viewer.html) includes a reference to the SheetJS library hosted on cdn.sheetjs.com. This is used to render Excel (.xlsx) files locally in the browser during result review. SheetJS is a well-known service for spreadsheet data processing.
[DATA_EXFILTRATION]: The skill records execution metrics, including token consumption and task duration, into local timing.json files within the iteration workspace. This data is collected solely for local performance benchmarking and variance analysis.
[DYNAMIC_EXECUTION]: The eval-viewer/generate_review.py script dynamically constructs a standalone HTML report by embedding local workspace file contents and benchmark statistics as a JSON payload into a template. This process occurs locally to provide the user with a qualitative review interface.

skill-creator