AWS MCP Server

An MCP server that provides tools for creating and managing AWS resources using boto3.

Features

Resource Creation

• create_ec2_instance: Launch EC2 instances
• create_s3_bucket: Create S3 buckets with versioning and encryption
• create_vpc: Create VPCs with DNS settings
• create_subnet: Create subnets in VPCs
• create_security_group: Create security groups with ingress rules
• create_rds_instance: Create RDS database instances

Resource Management

• list_resources: List resources by type (EC2, S3, VPC, subnets, RDS, security groups)
• describe_resource: Get detailed information about specific resources
• delete_resource: Delete resources (with safety checks)

Prerequisites

• Python 3.10+
• AWS CLI configured with credentials
• Appropriate AWS IAM permissions

Installation

cd aws-mcp-server
pip install -r requirements.txt

AWS Credentials

Configure AWS credentials using one of these methods:

1. AWS CLI: aws configure
2. Environment variables:

   export AWS_ACCESS_KEY_ID=your_key
   export AWS_SECRET_ACCESS_KEY=your_secret
   export AWS_DEFAULT_REGION=us-east-1
   

3. IAM role (if running on EC2)

Configuration

Add to your Claude Code settings (~/.claude/settings.json):

{
  "mcpServers": {
    "aws": {
      "command": "python",
      "args": ["/Users/youruser/Documents/Base/DevOps-ClaudeAi/test-cases/SM1/my-first-vpc/aws-mcp-server/server.py"]
    }
  }
}

Usage Examples

User: "Create a VPC with CIDR 10.0.0.0/16 in us-east-1"
Claude: [calls create_vpc]

User: "List all EC2 instances in us-west-2"
Claude: [calls list_resources with resource_type=ec2, region=us-west-2]

User: "Create an S3 bucket named my-app-bucket with versioning enabled"
Claude: [calls create_s3_bucket with enable_versioning=true]

User: "Create a t2.micro EC2 instance using ami-12345"
Claude: [calls create_ec2_instance]

Security Notes

⚠️ IMPORTANT:

• This server can create real AWS resources that incur costs
• Always review actions before confirming
• Use IAM policies to limit permissions (principle of least privilege)
• Consider using AWS Budget Alerts
• The delete_resource tool is intentionally restricted - use AWS Console for deletions

Recommended IAM Policy

Create a restricted IAM user/role with only the permissions you need:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "ec2:Describe*",
        "ec2:RunInstances",
        "ec2:CreateVpc",
        "ec2:CreateSubnet",
        "ec2:CreateSecurityGroup",
        "s3:CreateBucket",
        "s3:ListBucket",
        "s3:PutBucketVersioning",
        "rds:CreateDBInstance",
        "rds:DescribeDBInstances"
      ],
      "Resource": "*"
    }
  ]
}

Error Handling

The server provides detailed error messages from AWS API responses, including:

• Permission errors (Access Denied)
• Resource conflicts (already exists)
• Quota limits exceeded
• Invalid parameters