gh-dieshen-claude-marketplace-plugins-mcp-integrations/commands/setup-mcps.md

MCP Server Integration Guide

Complete guide for setting up Model Context Protocol (MCP) servers with Claude Code.

What are MCP Servers?

MCP servers provide Claude Code with direct access to external systems and data sources. They enable real-time interaction with databases, cloud providers, Docker/Kubernetes, and more.

Available MCP Servers

1. Database MCP

Connect to SQL and NoSQL databases for real-time queries and schema exploration.

2. Vector Database MCP

Integrate with Pinecone, Weaviate, or Qdrant for AI/ML workflows.

3. Cloud Provider MCPs

Direct integration with AWS, GCP, and Azure services.

4. Docker/Kubernetes MCP

Manage containers and clusters directly from Claude Code.

5. GitHub MCP

Enhanced Git operations, PR reviews, and issue management.

Installation

Prerequisites

# Node.js 18+ required
node --version

# Claude Code CLI
claude --version

General MCP Setup

1. Install MCP server package:

npm install -g @modelcontextprotocol/server-{name}

2. Configure in Claude Code settings:

{
  "mcpServers": {
    "server-name": {
      "command": "mcp-server-name",
      "args": ["--config", "path/to/config.json"],
      "env": {
        "API_KEY": "your-api-key"
      }
    }
  }
}

3. Restart Claude Code to load the MCP server

Database MCP Setup

PostgreSQL

# Install
npm install -g @modelcontextprotocol/server-postgres

# Configure (~/.claude/config.json)
{
  "mcpServers": {
    "postgres": {
      "command": "mcp-server-postgres",
      "args": [],
      "env": {
        "DATABASE_URL": "postgresql://user:pass@localhost:5432/dbname"
      }
    }
  }
}

Usage:

/query SELECT * FROM users WHERE email = 'test@example.com'
/schema users
/indexes users

MongoDB

# Install
npm install -g @modelcontextprotocol/server-mongodb

# Configure
{
  "mcpServers": {
    "mongodb": {
      "command": "mcp-server-mongodb",
      "env": {
        "MONGODB_URI": "mongodb://localhost:27017/mydb"
      }
    }
  }
}

Usage:

/find users {"email": "test@example.com"}
/aggregate users [{"$group": {"_id": "$role", "count": {"$sum": 1}}}]

Vector Database MCP Setup

Pinecone

# Install
npm install -g @modelcontextprotocol/server-pinecone

# Configure
{
  "mcpServers": {
    "pinecone": {
      "command": "mcp-server-pinecone",
      "env": {
        "PINECONE_API_KEY": "your-api-key",
        "PINECONE_ENVIRONMENT": "us-east-1-aws"
      }
    }
  }
}

Usage:

/vector-search "machine learning concepts" --top-k 5
/upsert-vectors [{"id": "vec1", "values": [...], "metadata": {...}}]
/delete-vectors ["vec1", "vec2"]

Cloud Provider MCP Setup

AWS

# Install
npm install -g @modelcontextprotocol/server-aws

# Configure
{
  "mcpServers": {
    "aws": {
      "command": "mcp-server-aws",
      "env": {
        "AWS_ACCESS_KEY_ID": "your-access-key",
        "AWS_SECRET_ACCESS_KEY": "your-secret-key",
        "AWS_REGION": "us-east-1"
      }
    }
  }
}

Usage:

/aws s3 ls s3://my-bucket
/aws ec2 describe-instances --filters "Name=tag:Environment,Values=production"
/aws lambda invoke my-function --payload '{"key": "value"}'
/aws rds describe-db-instances

GCP

# Install
npm install -g @modelcontextprotocol/server-gcp

# Configure
{
  "mcpServers": {
    "gcp": {
      "command": "mcp-server-gcp",
      "env": {
        "GOOGLE_APPLICATION_CREDENTIALS": "/path/to/service-account.json",
        "GCP_PROJECT_ID": "my-project"
      }
    }
  }
}

Usage:

/gcp compute instances list
/gcp storage buckets list
/gcp sql instances list

Docker/Kubernetes MCP Setup

Docker

# Install
npm install -g @modelcontextprotocol/server-docker

# Configure
{
  "mcpServers": {
    "docker": {
      "command": "mcp-server-docker",
      "args": ["--socket", "/var/run/docker.sock"]
    }
  }
}

Usage:

/docker ps
/docker images
/docker logs container-name --tail 100
/docker exec container-name ls -la
/docker stats

Kubernetes

# Install
npm install -g @modelcontextprotocol/server-kubernetes

# Configure
{
  "mcpServers": {
    "k8s": {
      "command": "mcp-server-kubernetes",
      "env": {
        "KUBECONFIG": "~/.kube/config"
      }
    }
  }
}

Usage:

/kubectl get pods -n production
/kubectl describe deployment myapp
/kubectl logs deployment/myapp --tail=100
/kubectl get services
/kubectl top pods

GitHub MCP Setup

# Install
npm install -g @modelcontextprotocol/server-github

# Configure
{
  "mcpServers": {
    "github": {
      "command": "mcp-server-github",
      "env": {
        "GITHUB_TOKEN": "ghp_your_token_here"
      }
    }
  }
}

Usage:

/gh pr list --repo owner/repo
/gh pr view 123
/gh pr review 123 --comment "LGTM"
/gh issue create --title "Bug" --body "Description"
/gh workflow run ci.yml

Security Best Practices

1. Store credentials securely:

# Use environment variables
export DATABASE_URL="postgresql://..."

# Or use a secrets manager
{
  "mcpServers": {
    "postgres": {
      "command": "mcp-server-postgres",
      "env": {
        "DATABASE_URL": "${env:DATABASE_URL}"
      }
    }
  }
}

2. Use read-only access when possible:

-- Create read-only database user
CREATE USER readonly_user WITH PASSWORD 'secure_password';
GRANT CONNECT ON DATABASE mydb TO readonly_user;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO readonly_user;

3. Limit permissions:

• Use least privilege principle
• Create service accounts with minimal permissions
• Regularly rotate credentials

4. Monitor usage:

• Log all MCP operations
• Set up alerts for unusual activity
• Review access logs regularly

Troubleshooting

MCP Server Not Starting

# Check if server is installed
which mcp-server-postgres

# Test server manually
mcp-server-postgres

# Check logs
tail -f ~/.claude/logs/mcp-server-name.log

Connection Issues

# Test database connection
psql postgresql://user:pass@localhost:5432/dbname

# Test API credentials
curl -H "Authorization: Bearer $API_KEY" https://api.provider.com/test

Permission Errors

• Verify credentials are correct
• Check IAM roles/permissions
• Ensure firewall allows connections
• Verify network connectivity

Creating Custom MCP Servers

Basic MCP Server Template

import { Server } from '@modelcontextprotocol/sdk/server/index.js';
import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';

const server = new Server(
  {
    name: 'my-custom-server',
    version: '1.0.0',
  },
  {
    capabilities: {
      tools: {},
    },
  }
);

// Define tools
server.setRequestHandler('tools/list', async () => {
  return {
    tools: [
      {
        name: 'my_tool',
        description: 'Description of what this tool does',
        inputSchema: {
          type: 'object',
          properties: {
            param1: {
              type: 'string',
              description: 'Parameter description',
            },
          },
          required: ['param1'],
        },
      },
    ],
  };
});

// Handle tool calls
server.setRequestHandler('tools/call', async (request) => {
  const { name, arguments: args } = request.params;

  if (name === 'my_tool') {
    const result = await doSomething(args.param1);
    return {
      content: [
        {
          type: 'text',
          text: JSON.stringify(result, null, 2),
        },
      ],
    };
  }

  throw new Error(`Unknown tool: ${name}`);
});

// Start server
const transport = new StdioServerTransport();
await server.connect(transport);

Best Practices

1. Test MCP servers locally before deploying
2. Use version control for MCP configurations
3. Document custom MCP servers
4. Monitor performance and errors
5. Keep MCP servers updated
6. Implement proper error handling
7. Use connection pooling for databases
8. Cache frequently accessed data
9. Set appropriate timeouts
10. Log operations for debugging

Example Workflows

Database Query & Analysis

1. /query SELECT * FROM users WHERE created_at > NOW() - INTERVAL '7 days'
2. Analyze results
3. /query UPDATE users SET status = 'active' WHERE ...
4. Verify changes

Cloud Deployment

1. /docker build -t myapp:latest .
2. /docker push myapp:latest
3. /kubectl set image deployment/myapp myapp=myapp:latest
4. /kubectl rollout status deployment/myapp

AI/ML Workflow

1. Generate embeddings for documents
2. /vector-search "relevant query" --top-k 10
3. Use results in RAG pipeline
4. /upsert-vectors with new documents