Hugging Face

Explore agent instructions for datasets, training, evaluation and Hub operations. Hugging Face’s task collection helps you identify the required tools and resources before starting an AI or ML workflow.

About Hugging Face

Hugging Face publishes its Skills in the huggingface GitHub organization. The repository describes AI and machine-learning tasks such as dataset creation, model training and evaluation, with instructions and supporting resources packaged per task.

The README distinguishes Hub operations through the hf-cli bootstrap path from additional workflow Skills. Dataset rights, model licenses, account access and compute requirements remain specific to each task; the publisher profile is not a grant of training capacity.

Explore Other Publishers

Meet other people and teams publishing open tools, projects, and Agent Skills.

Hugging Face FAQ

Where are the Hugging Face Skills maintained?

The upstream source is huggingface/skills, owned by the Hugging Face GitHub organization.

What distinguishes this collection from a coding-only pack?

Its tasks include datasets, model training and evaluation, alongside Hub operations and their supporting tools.

How do Hub bootstrap instructions differ from workflow Skills?

The README presents hf-cli as a bootstrap path for core Hub operations, while additional task workflows can be installed or discovered separately.

Does installing a Skill provide compute?

No. Training and evaluation still need an appropriate runtime, hardware or hosted job resources.

Do all models and datasets share the Skills license?

No. Review each model and dataset’s terms independently of the instruction repository.

Where should I follow workflow updates?

Read the Skills README and the individual task instructions alongside the relevant Hub documentation.

Follow Hugging Face on GitHub

Inspect the Hugging Face Skills catalog and the task-specific files before preparing a training or evaluation job.

Visit Hugging Face on GitHub