The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation provides standard installation commands using the uv package manager for transformers, huggingface_hub, datasets, evaluate, and other domain-specific libraries like timm and librosa.- [REMOTE_CODE_EXECUTION]: The skill documents the trust_remote_code parameter, which is a feature of the library that allows execution of custom logic from model repositories. Importantly, the skill includes a safety warning advising users to only use this feature when they have reviewed the source code from the repository.- [CREDENTIALS_SAFE]: The skill includes comprehensive guidance on managing Hugging Face authentication tokens. It advises against hardcoding secrets, suggests using the hf auth login CLI for secure storage, and recommends using environment variables via secret managers for CI/CD environments.

transformers