mode-manager-mcp 0.1.4

Mode Manager MCP: Server for managing VS Code .chatmode.md and .instructions.md files (GitHub Copilot prompts)

Mode Manager MCP

🧠 Personal AI Memory + VS Code Prompt Management

Streamlined MCP server that gives AI assistants a persistent memory of your preferences and manages VS Code .chatmode.md and .instructions.md files.

🎯 Primary Feature: AI Memory

One tool to rule them all: remember()

• 🧠 Personal AI Memory - Store preferences, facts, and context that AI should remember about you
• 🔄 Auto-setup - Creates memory file automatically in your VS Code prompts directory
• 💾 Persistent - Memories are saved as a VS Code instruction file that AI assistants can access
• ⚡ Simple - Just one tool instead of managing multiple complex operations

🧠 Primary Usage: Personal AI Memory

Store Information About Yourself

Ask Copilot: "Remember that I prefer Python over JavaScript for backend development"

Add Personal Context

Ask Copilot: "Remember that I work at Oatly and focus on sustainability projects"

Save Preferences

Ask Copilot: "Remember that I like detailed code comments and use Black for Python formatting"

How It Works

1. First use - Automatically creates memory.instructions.md in your VS Code prompts directory
2. Each memory - Adds timestamped entries to your personal memory file
3. AI access - Any AI assistant can read your memories when the instruction is active
4. Persistent - Memories survive across VS Code sessions and conversations

📚 The Mode Manager MCP Library

The official library contains 20 curated entries with proper license attribution:

Popular Chatmodes

• Beast Mode 3.1 - Advanced autonomous coding agent
• AI Architect - System design and architecture specialist
• Code Reviewer - Comprehensive code review assistant
• Tech Lead - Technical leadership and project management
• Security Expert - Cybersecurity and secure coding practices

Professional Instructions

• Python Developer - Modern Python patterns and best practices
• React Developer - Frontend development with React
• API Design Guidelines - RESTful API design principles
• Git Workflow Standards - Professional version control practices
• Testing Strategy - Comprehensive testing patterns

All entries include:

• 📄 License information (MIT, Apache 2.0, Open Source)
• 👤 Author attribution
• 🏷️ Categories and tags
• 🔗 Source links to original content

✨ Additional Features

• 📂 Manage VS Code prompt files - Create, edit, and organize chatmodes and instructions
• 📚 Browse the curated library - Install popular chatmodes and instructions from the community with license attribution
• 🔄 Keep files updated - Update chatmodes from source while preserving your custom tool settings
• 🛡️ Safe operations - Automatic backups before any deletions
• ⚙️ Configurable library - Use the official library or point to your own custom collection

🚀 Quick Setup

1. Install Hatch

Hatch is a modern Python project manager that handles virtual environments, dependencies, and build processes automatically.

pip install hatch

2. Clone and Setup

git clone https://github.com/NiclasOlofsson/node-manager-mcp.git
cd node-manager-mcp

Hatch will automatically create a virtual environment and install dependencies when you run commands.

3. Test it works

hatch run test

4. Add to VS Code

Copy the example configuration:

cp mcp-config-example.json .vscode/mcp.json

Or manually add to your VS Code MCP settings:

{
  "mcpServers": {
    "mode-manager": {
      "command": "python",
      "args": ["-m", "src.mode_manager_mcp"],
      "cwd": "."
    }
  }
}

Note: If you're developing on the project, you can also run the server in development mode:

# Test the server locally
hatch run python -m src.mode_manager_mcp

⚙️ Library Configuration

The server uses the official Mode Manager MCP Library by default, but you can configure it to use custom libraries for development or organizational needs.

Default Library

https://raw.githubusercontent.com/NiclasOlofsson/node-manager-mcp/refs/heads/main/library/memory-mode-library.json

Configuration Options

Command Line Parameter

python -m src.mode_manager_mcp --library-url "https://your-custom-library.json"

Environment Variable

export MCP_LIBRARY_URL="https://your-custom-library.json"
python -m src.mode_manager_mcp

VS Code Configuration

{
  "mcpServers": {
    "mode-manager": {
      "command": "python",
      "args": ["-m", "src.mode_manager_mcp", "--library-url", "https://your-custom-library.json"],
      "cwd": "."
    }
  }
}

Common Use Cases

Development/Testing:

# Use a local development library
python -m http.server 8000 --directory ./library
export MCP_LIBRARY_URL="http://localhost:8000/memory-mode-library.json"

Custom Organization Library:

# Point to your organization's curated library
export MCP_LIBRARY_URL="https://raw.githubusercontent.com/yourorg/custom-prompts/main/library.json"

💡 Usage Examples

Browse the Library

Ask Copilot: "What chatmodes are available in the library?"

Install a Popular Chatmode

Ask Copilot: "Install Beast Mode 3.1 from the library"

Create Your Own

Ask Copilot: "Create a new chatmode for API testing"

Manage Files

Ask Copilot: "List my current chatmodes" or "Update my chatmode from its source"

📁 Where Files Are Stored

Your prompt files live in VS Code's prompts directory:

• Windows: %APPDATA%\Code\User\prompts
• macOS: ~/Library/Application Support/Code/User/prompts
• Linux: ~/.config/Code/User/prompts

🤝 Contributing

Want to help improve Mode Manager MCP? Check out CONTRIBUTING.md for development setup, coding guidelines, and how to submit changes.

📄 License

MIT License - see LICENSE for details.