5-Minute Setup

Get Skillsmith running in 5 minutes. Choose your preferred workflow below.

Before You Begin

Choose Your Workflow

Using an MCP Client

Step 1: Open Your Project, Then Open Your AI Agent

Skillsmith reads your project's package.json and file structure to recommend relevant skills, so open your AI agent from inside the project you want skills for.

cd /path/to/your/project

"MCP client" just means your AI coding agent — Claude Code, Cursor, GitHub Copilot, Windsurf, and Codex CLI all work the same way here. Already have one open in this project? Skip to Step 2.

Installing Claude Code (skip this if you already have an agent open):

# Install via npm (recommended)
npm install -g @anthropic-ai/claude-code

# Or via Homebrew
brew install claude-code

# Install via npm (requires Node.js)
npm install -g @anthropic-ai/claude-code

# Or download the installer from:
# https://claude.ai/download

# Install via npm
npm install -g @anthropic-ai/claude-code

# Or via the install script
curl -fsSL https://claude.ai/install.sh | sh

Step 2: Add the Skillsmith MCP Server

Using Claude Code? Start with Option A. Using any other agent? Skip to Option B.

Option A — Claude Code's built-in command (fastest)

claude mcp add skillsmith -- npx -y @skillsmith/mcp-server

The -- separates Claude Code's own flags from the command it should run. This adds the server at the default --scope local (this project only); pass --scope user to make it available in every project instead, or --scope project to share it via a checked-in config. Verify with claude mcp list; remove it with claude mcp remove skillsmith.

Note: skillsmith here is just a local name you're choosing for this install — it doesn't need to match the @skillsmith/mcp-server package name used in Option B/C's config files.

Option B — Ask your agent in chat (works with any agent)

This is a normal file-edit request, not a special "auto-config" feature — your agent will edit a config file for you, and you should expect the usual permission prompt the first time it writes to a new file. Paste this into your agent's chat:

Please add the Skillsmith MCP server to my MCP client configuration.
Create the config file if it doesn't exist yet, and show me the diff
before you save. Use this entry:

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "npx",
      "args": ["-y", "@skillsmith/mcp-server"]
    }
  }
}

If you are NOT Claude Code, also add an "env" block to that entry with
SKILLSMITH_CLIENT set to your own client name (e.g. "windsurf" for Windsurf,
"copilot" for GitHub Copilot) so skills install to your own skills directory
instead of Claude Code's default. If you're not sure of the right value,
ask me before saving.

Tell me which file you changed when you're done.

Not using Claude Code? Skills install to your agent's own directory by default (e.g. ~/.claude/skills/ for Claude Code). The prompt above already asks your agent to set SKILLSMITH_CLIENT for itself when it isn't Claude Code — see the client-specific snippets in Option C below for exact examples if you'd rather edit the file yourself.

Using Cursor? Skip the prompt above and use Option C's Cursor snippet instead — Cursor's bundled Node cannot resolve packages via npx (a real ENOENT), so the "command": "npx" entry above will not connect. Option C's Cursor section gives you the resolved-path form that's confirmed to work.

Option C — Edit the config file yourself

Pick the snippet for your MCP client and add it to the matching config file. Every snippet uses the same shape — only the config-file path (and, for Codex/Hermes, the format) differs.

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "npx",
      "args": ["-y", "@skillsmith/mcp-server"]
    }
  }
}

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "<paste output of: which skillsmith-mcp (macOS/Linux) or where skillsmith-mcp (Windows)>",
      "env": {
        "SKILLSMITH_CLIENT": "cursor"
      }
    }
  }
}

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "npx",
      "args": ["-y", "@skillsmith/mcp-server"]
    }
  }
}

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "npx",
      "args": ["-y", "@skillsmith/mcp-server"]
    }
  }
}

[mcp_servers.@skillsmith/mcp-server]
command = "npx"
args = ["-y", "@skillsmith/mcp-server"]

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "npx",
      "args": ["-y", "@skillsmith/mcp-server"]
    }
  }
}

{
  "mcp": {
    "@skillsmith/mcp-server": {
      "type": "local",
      "command": ["npx", "-y", "@skillsmith/mcp-server"],
      "enabled": true
    }
  }
}

mcp_servers:
  @skillsmith/mcp-server:
    command: "npx"
    args: ["-y", "@skillsmith/mcp-server"]

[mcp_servers.@skillsmith/mcp-server]
command = "npx"
args = ["-y", "@skillsmith/mcp-server"]

{
  "mcpServers": {
    "@skillsmith/mcp-server": {
      "command": "npx",
      "args": ["-y", "@skillsmith/mcp-server"],
      "env": {
        "SKILLSMITH_CLIENT": "antigravity"
      }
    }
  }
}

After any of the three options above, restart your agent to load Skillsmith.

Step 3: Discover Skills for Your Project

You're already in your project with your agent open from Step 1 — just ask:

Using the CLI Directly

Step 1: Install the Skillsmith CLI

# Install globally
npm install -g @skillsmith/cli

# Verify installation
skillsmith --version

# Run without installing (downloads on first use)
npx @skillsmith/cli search testing

# Short alias
npx sklx search testing

Step 2: Navigate to Your Project

Open your terminal and navigate to your project directory:

# Navigate to your project
cd ~/projects/my-react-app

# Verify you're in the right place
ls package.json

# Navigate to your project
cd C:\Users\YourName\projects\my-react-app

# Verify you're in the right place
dir package.json

Why navigate here? Skillsmith can analyze your project's package.json, dependencies, and file structure to recommend relevant skills.

Step 3: Search for Skills

# Search by keyword
skillsmith search testing

# Search with filters
skillsmith search react --category development
skillsmith search git --tier verified

# Get recommendations for current project
skillsmith recommend

The recommend command analyzes your project and suggests relevant skills.

Step 4: Install Skills

# Install by skill name
skillsmith install jest-helper

# Install by full ID
skillsmith install community/react-patterns

# Install multiple skills
skillsmith install jest-helper commit react-patterns

Step 5: Verify Installation

# List installed skills
skillsmith list

# Show skill details
skillsmith info jest-helper

Next Steps