ProduktEntdecker/infrastructure-mcp-server

Infrastructure MCP Server

Universal Model Context Protocol (MCP) server that provides AI agents with standardized access to development infrastructure, tools, and scripts.

🚀 Transform fragmented AI agent tooling into a unified infrastructure accessible by any MCP-enabled agent.

Features

• ✅ Universal Script Access: startup.sh and wrapup.sh available to all MCP-enabled agents
• ✅ Tools Inventory: Complete access to 130+ tool inventory with filtering
• ✅ Session Management: Session status and context awareness
• ✅ Help System: Built-in help for all scripts and tools

Available Tools

Scripts

• startup - Execute universal startup script with project detection and environment setup
• wrapup - Execute universal wrapup script with PR creation and session handover
• startup_help - Get detailed help for startup script
• wrapup_help - Get detailed help for wrapup script

Session Management

• session_status - Get current development session status and context

Tools Discovery

• tools_inventory - Get complete inventory of 130+ available tools with filtering options

Installation & Setup

1. Clone and Build the Server

# Clone the repository
git clone https://github.com/ProduktEntdecker/infrastructure-mcp-server.git
cd infrastructure-mcp-server

# Install dependencies and build
npm install
npm run build

2. Configure Claude Code

Add to your project's .mcp.json:

{
  "mcpServers": {
    "infrastructure": {
      "command": "node",
      "args": ["/Users/floriansteiner/Documents/GitHub/infrastructure-mcp-server/dist/server.js"],
      "description": "Universal infrastructure tools and scripts access"
    }
  }
}

Enable project MCP servers in ~/.claude/settings.local.json:

{
  "enableAllProjectMcpServers": true
}

3. Usage Examples

Execute Startup Script

// MCP Tool Call
{
  "name": "startup",
  "arguments": {
    "check": true,  // Check-only mode
    "fresh": false  // Not a fresh start
  }
}

Get Tools Inventory

// MCP Tool Call  
{
  "name": "tools_inventory",
  "arguments": {
    "category": "AI Development",
    "mcp": true,
    "search": "github"
  }
}

Execute Wrapup with PR Creation

// MCP Tool Call
{
  "name": "wrapup", 
  "arguments": {
    "branch": "feature/mcp-integration",
    "title": "Add universal MCP infrastructure access"
  }
}

Architecture

AI Agents (Claude, Codex, Warp) 
    ↓ MCP Protocol
Infrastructure MCP Server
    ↓ Shell Execution
startup.sh / wrapup.sh / TOOLS.md

Benefits for AI Agents

Before MCP Integration 🚫

• No access to startup/wrapup scripts
• Manual context provision for each agent
• Inconsistent security protocol adherence
• Tool discovery required for each session

After MCP Integration ✅

• Universal script access via MCP protocol
• Automatic context and tool discovery
• Consistent security framework enforcement
• Session continuity via HANDOVER.md

Development

Run in Development Mode

npm run dev

Build for Production

npm run build
npm start

Test the Server

# Test if server starts correctly
timeout 5s npm run dev || echo "Started successfully"

Next Steps

1. GitHub Codex Integration - Add .github/copilot-mcp.yml
2. Warp AI Integration - Configure ~/.warp/mcp-config.json
3. Enhanced Security Services - Add NPM audit and secret scanning
4. Workflow Services - Add CodeRabbit integration and deployment checks
5. Distribution - Publish as npm package for easy installation

This MCP server transforms fragmented tool access into a unified infrastructure that any MCP-enabled AI agent can leverage immediately.