MCP Automation Hub

Overview

This Skill provides CLI and automation tools for managing your MCP (Model Context Protocol) server, n8n workflow integration, and service monitoring.

Quick Start

Health Check

Verify all MCP services are running:

curl http://localhost:3000/health

Expected response: {"status": "healthy"} with 200 status code.

Deploy to n8n

Sync local custom nodes to your n8n instance:

cd ~/mcp-automation-hub && pnpm deploy:n8n

View Logs

Check recent MCP server activity:

tail -n 50 /tmp/mcp-automation.out.log

For error logs:

tail -n 50 /tmp/mcp-automation.err.log

Service Management

Start MCP Server

launchctl load ~/Library/LaunchAgents/com.mcp.automation.plist

Stop MCP Server

launchctl unload ~/Library/LaunchAgents/com.mcp.automation.plist

Restart MCP Server

launchctl unload ~/Library/LaunchAgents/com.mcp.automation.plist
launchctl load ~/Library/LaunchAgents/com.mcp.automation.plist

Check Service Status

launchctl list | grep com.mcp.automation

Configuration

Environment Setup

Configuration file: .env.example

Copy and customize:

cp .env.example .env

Key variables:

• •MCP_PORT: Server port (default: 3000)
• •N8N_API_URL: n8n instance URL
• •N8N_API_KEY: n8n API authentication key

macOS Shortcuts Integration

For details on Apple Shortcuts automation, see Shortcuts.md.

Key shortcuts:

1. •MCP Health Check - Quick service verification
2. •Deploy to n8n - One-click deployment
3. •MCP Logs - Instant log viewing

Scripts

Utility scripts are in the scripts/ directory:

• •scripts/setup-mac.sh - Initial Mac environment setup
• •Additional automation scripts as needed

Troubleshooting

Service Won't Start

1. •

   Check if port 3000 is available:

   bash

   lsof -i :3000
   

2. •

   Verify plist file exists:

   bash

   ls -la ~/Library/LaunchAgents/com.mcp.automation.plist
   

3. •

   Check error logs:

   bash

   cat /tmp/mcp-automation.err.log
   

n8n Deployment Fails

1. •Verify n8n is running and accessible
2. •Check N8N_API_KEY in .env
3. •Ensure dependencies are installed:
   bash

   pnpm install
   

Health Check Returns Error

1. •

   Confirm server is running:

   bash

   launchctl list | grep com.mcp.automation
   

2. •

   Test with verbose output:

   bash

   curl -v http://localhost:3000/health
   

3. •

   If server is down, restart it using the commands in Service Management section

Common Workflows

Deploy Changes to n8n

Copy this workflow and check off steps as you complete them:

Deployment Progress:
- [ ] Step 1: Make changes to n8n nodes
- [ ] Step 2: Test locally if possible
- [ ] Step 3: Run deployment command
- [ ] Step 4: Verify deployment in n8n UI
- [ ] Step 5: Test workflows using deployed nodes

Step 1: Edit custom node files in your project

Step 2: If you have a local n8n instance, test there first

Step 3: Deploy using:

cd ~/mcp-automation-hub && pnpm deploy:n8n

Step 4: Open n8n and confirm nodes appear in the node palette

Step 5: Create or run a workflow that uses your custom nodes

Monitor Service Health

1. •

   Run health check:

   bash

   curl http://localhost:3000/health
   

2. •

   If status is not 200:

   • •Check logs for errors
   • •Restart service
   • •Verify configuration
3. •

   Set up periodic monitoring (optional):

   bash

   # Add to crontab for hourly checks
   0 * * * * curl -s http://localhost:3000/health || osascript -e 'display notification "MCP Server is down!" with title "MCP Alert"'
   

Project Structure

mcp-automation-hub/
├── SKILL.md              # This file
├── Shortcuts.md          # Apple Shortcuts guide
├── SETUP.md             # Mac environment setup
├── README.md            # Project overview
├── scripts/             # Utility scripts
├── mcp-config.json      # MCP server config
├── package.json         # Dependencies and scripts
└── .env.example         # Configuration template

Dependencies

Required packages (auto-installed via pnpm install):

• •Node.js (v18+)
• •pnpm (package manager)
• •n8n (workflow automation)

Optional:

• •curl (for API testing)
• •jq (for JSON parsing)

Best Practices

1. •

   Always check health before deployments

   bash

   curl http://localhost:3000/health && pnpm deploy:n8n
   

2. •

   Keep logs clean - Rotate logs periodically:

   bash

   > /tmp/mcp-automation.out.log
   > /tmp/mcp-automation.err.log
   

3. •

   Use environment variables - Never hardcode credentials

4. •

   Test locally first - Verify changes before deploying to production n8n

5. •

   Monitor regularly - Set up automated health checks or use the Shortcuts

Additional Resources

For advanced MCP configuration, see the official MCP documentation.

For n8n custom node development, refer to n8n's developer documentation.