Start here4 min read

Get up and running in 30 seconds

Your AI coding assistant is only as good as the instructions you give it. Skills are those instructions — packaged, versioned, and installable in one command.

1Install your first skill

No setup, no config files. Just this:

bash
npx @skills-hub-ai/cli install code-review

This downloads the skill's SKILL.md file to ~/.claude/skills/code-review/ (or ~/.cursor/skills/ for Cursor with --target cursor).

2Use it

Type /code-review in Claude Code or Cursor. Done. Your AI now knows exactly how to review code the way you want it to. No prompt engineering, no copy-pasting instructions every session.

If a skill is a composition (a pipeline of multiple skills), all dependencies are installed automatically. Installing a "full story lifecycle" skill that chains /arch-review → /story-implementer → /pr brings all three child skills along.

3Works everywhere

Skills aren't locked to one tool. The skills-hub MCP server exposes your installed skills as prompts in any MCP-compatible AI tool — including Windsurf, Copilot, and others.

Claude Code

bash
claude mcp add skills-hub -- npx @skills-hub-ai/mcp

Cursor

Add to .cursor/mcp.json:

json
{
  "mcpServers": {
    "skills-hub": {
      "command": "npx",
      "args": ["@skills-hub-ai/mcp"]
    }
  }
}

Any other MCP client

Point your tool at npx @skills-hub-ai/mcp as a stdio MCP server. The server scans both ~/.claude/skills/ and ~/.cursor/skills/ and serves each installed skill as an MCP prompt.

4Browse and search

The skill catalog has thousands of skills from Anthropic, Microsoft, Google, Cloudflare, and the developer community. Search by keyword, natural language, or filter by company, category, and platform. Every skill shows quality scores, install counts, and community ratings.

Or search from your terminal:

bash
npx @skills-hub-ai/cli search "code review"
npx @skills-hub-ai/cli search --category review --sort highest_rated

5Publish your own

Got a workflow that works? Package it. A skill is just a SKILL.md file with YAML frontmatter:

markdown
---
name: My Skill
description: What this skill does in one sentence
version: 1.0.0
category: code-quality
platforms:
  - CLAUDE_CODE
---

Your skill instructions go here. This is the content
that will be injected into the AI assistant's context.

Then publish via the CLI or web:

bash
npx @skills-hub-ai/cli login          # sign in (GitHub, Google, or email)
npx @skills-hub-ai/cli publish        # publishes ./SKILL.md

Prefer the web? Use the publish page — describe what you want and AI generates the name, description, category, tags, and instructions for you.

6Guided setup

Run init for a guided walkthrough that checks your platform, auth status, installed skills, and suggests popular ones:

bash
npx @skills-hub-ai/cli init

What's next