The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a standalone Python helper script (scripts/new_experiment.py) designed to generate Markdown stubs for new experiments. The script uses standard libraries (argparse, pathlib) and employs safe file handling practices, such as using the x mode in open() to ensure files are created exclusively without risk of overwriting existing data. It includes input validation for the experiment slug using a strict regular expression to prevent path traversal or malformed identifiers.
[SAFE]: The documentation establishes a rigorous audit trail for modeling research, requiring the pre-declaration of hypotheses, success metrics, and data snapshots. These practices align with security and integrity best practices for data science workflows.
[SAFE]: No external network dependencies, remote code downloads, or credential-handling patterns were detected within the skill or its associated scripts.

experiment-log