Hugging Face Skills
Official Hugging Face Agent Skills for working with models, datasets, and Spaces.
About
Hugging Face Skills is the official Agent Skills repository from Hugging Face, containing around 26 SKILL.md files that teach coding agents how to work with the Hugging Face ecosystem. The skills cover practical workflows around models, datasets, transformers, and Spaces, packaged in the open Agent Skills format so they install into Claude Code, Codex CLI, Cursor, and other compatible editors.
The collection was confirmed via GitHub git-tree inspection in a July 2026 audit of vendor skill repos. As a first-party resource from the central hub of open ML, it is the authoritative way to give an agent grounded knowledge of Hugging Face APIs and conventions rather than relying on the model's possibly dated training data. Most useful for ML practitioners building on the Hub.
What it can do
- 26 first-party skills for the Hugging Face ecosystem
- Guidance for models, datasets, and Spaces workflows
- Open SKILL.md format for all major coding agents
- Maintained by Hugging Face
Setup
Install via npx skills add huggingface/skills or add the repo as a plugin marketplace.
Prerequisites
A skill-aware coding agent; a Hugging Face account for API-backed workflows.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 75; Trending score 80/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.