NSAF MCP Server

This is a Model Context Protocol (MCP) server for the Neuro-Symbolic Autonomy Framework (NSAF). It allows AI assistants to interact with the NSAF framework through the MCP protocol.

Note: This repository includes both the NSAF framework code and the MCP server implementation, making it a complete package that can be deployed and used anywhere.

Note: This implementation uses a simplified version of the MCP protocol that doesn't require the official MCP SDK. It implements the core functionality needed to expose NSAF capabilities to AI assistants.

Features

• Run NSAF evolution with customizable parameters
• Compare different NSAF agent architectures
• Integrate NSAF capabilities into AI assistants

Prerequisites

• Node.js 18+ and npm
• Python 3.8+ with the NSAF framework installed

Installation

1. Clone this repository:

git clone https://github.com/ariunbolor/nsaf-mcp-server.git
cd nsaf-mcp-server

2. Install dependencies:

npm install

3. Build the server:

npm run build

Configuration

The server includes the NSAF framework code, so no additional configuration is required for basic usage. The MCP server is designed to work out-of-the-box when installed globally.

Usage

Running the server locally

npm start

Deploying to GitHub

1. Create a new GitHub repository for your MCP server:

   • Go to GitHub and create a new repository named nsaf-mcp-server
   • Initialize it with a README file

2. Use the provided setup script to push your code to GitHub:

# For a new repository
./setup-github-fixed.sh yourusername

# If the repository already exists and you want to overwrite its content
./setup-github-fixed.sh yourusername --force

The script will:

• Initialize git if needed
• Set up the remote repository
• Commit your changes
• Try to push to GitHub (with options to handle existing repositories)

3. Configure GitHub Actions for CI/CD (optional):
   • Create a .github/workflows directory
   • Add a workflow file for testing and building the server

Using with AI Assistants

To use this MCP server with AI assistants like Claude, you need to:

1. Install the server:

   Option 1: Install from GitHub (after pushing your code):

   npm install -g yourusername/nsaf-mcp-server

   Option 2: Install from your local directory:

   # Navigate to the nsaf-mcp-server directory
   cd nsaf_mcp_server
      
   # Install dependencies and build
   npm install
   npm run build
      
   # Install globally from the local directory
   npm install -g .

2. Add the server to your MCP settings configuration:

For Claude Desktop app, edit ~/Library/Application Support/Claude/claude_desktop_config.json (on macOS):

{
  "mcpServers": {
    "nsaf": {
      "command": "nsaf-mcp-server",
      "args": [],
      "env": {},
      "disabled": false,
      "autoApprove": []
    }
  }
}

For Cline, edit /Users/onthego/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json:

{
  "mcpServers": {
    "nsaf": {
      "command": "nsaf-mcp-server",
      "args": [],
      "env": {},
      "disabled": false,
      "autoApprove": []
    }
  }
}

Available Tools

run_nsaf_evolution

Run NSAF evolution with specified parameters.

Parameters:

• population_size: Size of the agent population (default: 20)
• generations: Number of generations to evolve (default: 10)
• mutation_rate: Mutation rate (0.0-1.0) (default: 0.2)
• crossover_rate: Crossover rate (0.0-1.0) (default: 0.7)
• architecture_complexity: Complexity of the agent architecture ('simple', 'medium', 'complex') (default: 'medium')

compare_nsaf_agents

Compare different NSAF agent architectures.

Parameters:

• architectures: List of architectures to compare (default: ['simple', 'medium', 'complex'])

License

MIT