Fix Docker containerization for MCP server

[Govinda-Fichtner]

Problem

The Heroku MCP Server fails to run properly in Docker containers due to several configuration issues:

1. Missing Heroku CLI: The container doesn't include the Heroku CLI, which is required for the MCP server to function
2. Incorrect Entrypoint: Uses generic dist/index.js instead of the proper bin/heroku-mcp-server.mjs entry point
3. File Permission Issues: Executable files don't have proper permissions in the container
4. Incomplete File Copying: Missing the bin/ directory which contains the actual server entry point

Solution

This PR includes essential fixes for Docker containerization:

Dockerfile Improvements

• Install Heroku CLI: Added proper Heroku CLI installation using the official installer
• Correct Entrypoint: Changed to use node bin/heroku-mcp-server.mjs as the entrypoint
• File Permissions: Ensure all executable files have proper permissions with chmod +x
• Complete File Structure: Copy both dist/ and bin/ directories to maintain the proper server structure

Docker Usage Documentation

• Comprehensive DOCKER.md: Added complete documentation for Docker usage
• Claude Desktop Integration: Includes configuration examples for Claude Desktop
• Cursor Integration: Includes configuration examples for Cursor
• Testing Instructions: Provides manual testing commands and expected responses
• Troubleshooting Guide: Common issues and solutions for Docker deployment

Testing

The fixed Docker container now properly:

• ✅ Responds to MCP protocol handshake requests
• ✅ Authenticates with Heroku API using environment variables
• ✅ Provides all expected MCP tools for Heroku platform management
• ✅ Works seamlessly with Claude Desktop and Cursor MCP integrations

Test Command

echo '{"jsonrpc": "2.0", "id": 1, "method": "initialize", "params": {"protocolVersion": "2024-11-05", "capabilities": {}, "clientInfo": {"name": "test", "version": "1.0.0"}}}' | docker run --rm -i --env-file .env heroku-mcp-server:latest

Configuration Examples

Claude Desktop

{
  "mcpServers": {
    "heroku": {
      "command": "docker",
      "args": ["run", "--rm", "-i", "--env-file", "/path/to/.env", "heroku-mcp-server:latest"]
    }
  }
}

Cursor

{
  "heroku": {
    "command": "docker",
    "args": ["run", "--rm", "-i", "--env-file", "/path/to/.env", "heroku-mcp-server:latest"]
  }
}

This enables the Heroku MCP Server to be easily deployed and used in containerized environments for AI coding assistants that support the Model Context Protocol.