freemcplab / Directory / Playground / Hugging Face Upskill

● Official huggingface ⚡ Instant

Hugging Face Upskill

Name: Hugging Face Upskill (Claude Skill)
Author: huggingface

Generate and evaluate agent skills for Claude Code, OpenCode, OpenAI Codex from datasets.

Hugging Face Upskill is a skill that generate and evaluate agent skills for claude code, opencode, openai codex from datasets Use it from Claude Code, Cursor, Codex, or any MCP-compatible / skills-compatible agent.

▶ Watch demo 📦 Install 💡 Use Cases

Why use it

Key features

Single-purpose integration
Compatible with Claude Code and other agents
Open-source on GitHub
Maintained by the publisher
Easy install via standard tooling

Live Demo

What it looks like in practice

huggingface-upskill.replay ▶ ready

0/0

Install

Pick your client

~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "huggingface-upskill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/huggingface/upskill",
        "~/.claude/tools/upskill"
      ],
      "_inferred": true
    }
  }
}

Open Claude Desktop → Settings → Developer → Edit Config. Restart after saving.

~/.cursor/mcp.json · .cursor/mcp.json

{
  "mcpServers": {
    "huggingface-upskill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/huggingface/upskill",
        "~/.claude/tools/upskill"
      ],
      "_inferred": true
    }
  }
}

Cursor uses the same mcpServers schema as Claude Desktop. Project config wins over global.

VS Code → Cline → MCP Servers → Edit

{
  "mcpServers": {
    "huggingface-upskill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/huggingface/upskill",
        "~/.claude/tools/upskill"
      ],
      "_inferred": true
    }
  }
}

Click the MCP Servers icon in the Cline sidebar, then "Edit Configuration".

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "huggingface-upskill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/huggingface/upskill",
        "~/.claude/tools/upskill"
      ],
      "_inferred": true
    }
  }
}

Same shape as Claude Desktop. Restart Windsurf to pick up changes.

~/.continue/config.json

{
  "mcpServers": [
    {
      "name": "huggingface-upskill",
      "command": "git",
      "args": [
        "clone",
        "https://github.com/huggingface/upskill",
        "~/.claude/tools/upskill"
      ]
    }
  ]
}

Continue uses an array of server objects rather than a map.

~/.config/zed/settings.json

{
  "context_servers": {
    "huggingface-upskill": {
      "command": {
        "path": "git",
        "args": [
          "clone",
          "https://github.com/huggingface/upskill",
          "~/.claude/tools/upskill"
        ]
      }
    }
  }
}

Add to context_servers. Zed hot-reloads on save.

claude mcp add huggingface-upskill -- git clone https://github.com/huggingface/upskill ~/.claude/tools/upskill

One-liner. Verify with claude mcp list. Remove with claude mcp remove.

Use Cases

Real-world ways to use Hugging Face Upskill

Use Hugging Face Upskill for its core workflow

👤 Engineers and power users ⏱ ~15 min intermediate

When to use: You want skill integration for skill work without writing glue code.

Prerequisites

Repo installed / cloned — See repo README

Flow

Install

Set up the Hugging Face Upskill skill.✓ Copied

→ skill listed in your agent config
Invoke

Use Hugging Face Upskill to run a representative task in skill.✓ Copied

→ Tool call succeeds; result returned

Outcome: Real skill work completed via Hugging Face Upskill with one chat.

Pitfalls

Auth/credential setup is the most common blocker. — Follow the repo README for env vars and tokens.

Combine with: filesystem

Combinations

Pair with other MCPs for X10 leverage

huggingface-upskill + filesystem

Pair with filesystem for skill workflows that need local file IO.

Use Hugging Face Upskill together with filesystem.✓ Copied

Tools

What this MCP exposes

Tool	Inputs	When to call	Cost
primary_call	task-specific	For the main use case	Varies

Cost & Limits

What this costs to run

API quota: Provider rate limits apply
Tokens per call: Varies
Monetary: Free / pay provider directly
Tip: Cache results; batch calls.

Security

Permissions, secrets, blast radius

Credential storage: Use environment variables or your OS keychain.

Data egress: Depends on integration

Troubleshooting

Common errors and fixes

Confirm install and restart your agent.

Resources

📖 Read the official README on GitHub

🐙 Browse open issues

🔍 Browse all 400+ MCP servers and Skills