scikit-learn

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing standard, well-known Python packages including scikit-learn, pandas, numpy, matplotlib, seaborn, imbalanced-learn, and category-encoders using the uv package manager. These are reputable libraries in the data science ecosystem.
  • [COMMAND_EXECUTION]: Includes example shell commands for package installation and Python scripts for executing classification and clustering pipelines. The operations are consistent with the skill's stated purpose.
  • [REMOTE_CODE_EXECUTION]: Provides documentation on model persistence using pickle and joblib. While these are the industry-standard methods for saving and loading scikit-learn models, they are inherently unsafe when used with untrusted files as they can execute arbitrary code during the deserialization process. The skill documents these as part of normal library usage.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 10:21 AM