The Agent Skills Directory

[SAFE]: The skill provides standard machine learning implementation patterns using trusted libraries such as PyTorch and Ultralytics. All external references point to official documentation (PyTorch, Ultralytics, Albumentations) or well-known model hubs (HuggingFace, Meta Research). All code samples follow secure practices, including proper data normalization and clear separation between training and validation stages. No instances of obfuscation, data exfiltration, or unauthorized command execution were found.

computer-vision