My Prompt MCP

中文版本 | English

A powerful MCP (Model Context Protocol) server that helps users quickly reuse their prompts and SOP workflows across different AI agents, providing seamless access to curated prompt templates and workflow guidance.

✨ Features

• 🎯 Rich Prompt Library: Access to carefully curated prompt workflows
• 🔧 Multiple Installation Methods: Support for npx, global installation, and source builds
• 🌐 Multi-IDE Support: Compatible with Claude Desktop, Cursor, Claude Code, Trae, Kiro, and Windsurf
• 🔒 Secure Authentication: Token-based authentication system
• ⚡ High Performance: Optimized for fast response times
• 🛠️ Easy Configuration: Simple setup with environment variables or command-line arguments

📦 Installation

Method 1: Using npx (Recommended)

npx my-prompt-mcp --token your_package_token_here

Method 2: Global Installation

npm install -g my-prompt-mcp
my-prompt-mcp --token your_package_token_here

Method 3: Build from Source

For developers who want to build from source, please refer to the Developer Guide.

⚙️ Configuration

Environment Variables

• MY_PROMPT_MCP_PACKAGE_TOKEN: Your package access token (required)
• MY_PROMPT_MCP_API_URL: API endpoint URL (optional, defaults to https://mypromptmcp.com)

Getting Your Token

1. Visit https://mypromptmcp.com
2. Register or log in to your account
3. Navigate to the API section
4. Generate a new access token
5. Copy the token for use in configuration

🛠️ Usage

Configuring MCP Server in Different IDEs

Claude Desktop

Add the following to your Claude Desktop configuration file:

Using npx (Recommended)

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Using Global Installation

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "my-prompt-mcp",
      "args": [
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Cursor IDE

There are two ways to configure MCP server in Cursor:

Method 1: Global Configuration Open Cursor settings (Ctrl/Cmd + Shift + P → "Cursor Settings"), add in MCP options:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Method 2: Project-level Configuration Create .cursor/mcp.json file in your project root:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Claude Code

In Claude Code, you can configure through the following methods:

Method 1: Using CLI Wizard

# Run in your project directory
claude-code /mcp
# Follow the wizard prompts to configure

Method 2: Direct Configuration File Editing Edit the .claude.json file:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Trae AI IDE

To configure MCP server in Trae Editor:

1. Click the "Add MCP Servers" button
2. Add the following configuration:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Kiro IDE

To configure MCP server in Kiro:

Create or edit the mcp.json configuration file:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Note: Do not commit mcp.json files containing sensitive tokens to Git repositories.

Windsurf IDE

To configure MCP server in Windsurf:

1. Open Windsurf settings page
2. Go to Cascade section
3. Scroll to MCP Servers section
4. Edit the ~/.codeium/windsurf/mcp_config.json file:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": [
        "my-prompt-mcp",
        "--token",
        "your_package_token_here"
      ]
    }
  }
}

Environment Variable Configuration (Optional)

Instead of specifying the token in command-line arguments, you can also use environment variables:

# Set environment variables
export MY_PROMPT_MCP_PACKAGE_TOKEN="your_package_token_here"
export MY_PROMPT_MCP_API_URL="https://mypromptmcp.com"  # Optional

Then omit the token parameter in configuration:

{
  "mcpServers": {
    "my-prompt-mcp": {
      "command": "npx",
      "args": ["my-prompt-mcp"],
      "env": {
        "MY_PROMPT_MCP_PACKAGE_TOKEN": "your_package_token_here"
      }
    }
  }
}

🔧 Available Tools

• get_prompt_workflows: Retrieve available prompt workflows
• execute_workflow: Execute a specific workflow with parameters
• list_categories: List all available workflow categories
• search_workflows: Search workflows by keywords
• get_workflow_details: Get detailed information about a specific workflow

⚠️ Important Notes

• STDIO Communication: This MCP server uses standard input/output for communication. Do not add any console.log or print statements in the code, as they will interfere with the MCP protocol.
• Workflow Execution: All workflows are executed in a sandboxed environment to ensure security and stability.

👨‍💻 Developer Guide

For developers who want to contribute to this project or build from source, please refer to the Developer Guide.

📄 License

MIT License

🤝 Contributing

We welcome contributions! Please see our Developer Guide for detailed information on how to contribute to this project.

Note: This is an MCP server designed to work with MCP-compatible clients. Make sure your client supports the MCP protocol before using this server.