skills
Give your agents the power of the Hugging Face ecosystem
Hugging Face Skills is a collection of AI agent skill definitions that extend coding agents with capabilities across the Hugging Face ecosystem. Each skill is a self-contained folder packaging instructions, scripts, and resources that teach agents how to perform specific ML/AI tasks like dataset creation, model training, deployment to SageMaker, and Hub operations. The skills follow the standardized Agent Skills format and work with Claude Code, Codex, Gemini CLI, and Cursor. The hf-cli skill serves as the recommended bootstrap, giving agents access to every Hugging Face Hub command, while 25+ additional specialized skills can be installed on demand for workflows like fine-tuning vision models, building Gradio demos, or publishing research papers.
Key Features
Use Cases
- 01Teaching coding agents to search, download, and upload models and datasets on Hugging Face Hub
- 02Automating SageMaker model deployment with proper IAM roles, container images, and CloudWatch monitoring
- 03Fine-tuning language models using TRL or Unsloth on Hugging Face Jobs cloud infrastructure
- 04Building and publishing Gradio demo applications on Hugging Face Spaces with ZeroGPU support
- 05Running local model evaluations with inspect-ai and lighteval frameworks
- 06Publishing and indexing research papers on Hugging Face with linked models and datasets
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
skills — FAQ
What is Hugging Face Skills?+
Hugging Face Skills is a repository of agent skill definitions that extend AI coding agents with capabilities for ML/AI tasks across the Hugging Face ecosystem. Each skill packages instructions, scripts, and resources in the standardized Agent Skills format with a SKILL.md file that agents parse to learn new workflows.
How do I install Hugging Face skills in Claude Code?+
First register the repository as a plugin marketplace with '/plugin marketplace add huggingface/skills', then install the CLI skill with '/plugin install hf-cli@huggingface/skills'. Additional skills can be installed using 'hf skills add <skill-name>' from the CLI.
Which AI coding agents support Hugging Face skills?+
Hugging Face skills work with Claude Code, Codex, Gemini CLI, and Cursor. They follow the standardized Agent Skills format, so any skill-aware coding agent can discover and load the SKILL.md instructions.
Do I need API keys or credentials to use these skills?+
Many skills require a Hugging Face account and CLI authentication (via 'huggingface-cli login'). SageMaker deployment skills require AWS credentials and appropriate IAM permissions. Local evaluation and training skills may only need local compute resources.
Are Hugging Face skills free to use?+
The skills themselves are open source and free to install. However, some operations they enable (like training on Hugging Face Jobs cloud GPUs or deploying to SageMaker) may incur cloud compute costs depending on your usage.
How do I use a skill after installing it?+
Once installed, mention the skill directly in your instructions to the coding agent, for example 'Use the HF LLM trainer skill to estimate GPU memory for a 70B model'. The agent automatically loads the corresponding SKILL.md instructions and helper scripts to complete the task.
How do I install skills?+
Open the source repository on GitHub and follow its README. skills is a skill — MCP Agents Market links you directly to the official repo.
Is skills free?+
skills is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.