MCP Server for AWS Resources

# AWS MCP Server - SyntropAI Ecosystem **Part of the [SyntropAI MCP Ecosystem](https://github.com/paihari/documentation-syntropai)** - A unified multi-cloud abstraction framework. This MCP (Model Context Protocol) server provides secure, dynamic access to AWS services through the innovative SyntropAI abstraction layer. Unlike traditional hardcoded service catalogs, this server supports **any AWS service** through dynamic SDK access with built-in security sandboxing. ## 🚀 Key Features - **Universal AWS Access**: Dynamic access to all AWS services without hardcoded limitations - **Secure Code Execution**: AST-based validation and sandboxed execution environment - **Provider-Agnostic Design**: Built on SyntropAI's unified abstraction pattern - **Future-Proof Architecture**: Automatically supports new AWS services without updates - **Docker Containerization**: Production-ready deployment ## 🏗️ Architecture This server implements the SyntropAI abstraction pattern: ``` Claude Desktop → MCP Protocol → AWS MCP Server → SyntropAIBox → boto3 → AWS Services ``` ### Core Components: - **AWSSession**: Unified AWS credential management using `BaseSession` - **AWSResourceQuerier**: Secure query execution extending `BaseQuerier` - **AST Sandbox**: Safe code execution with timeout protection - **Dynamic Schema**: Runtime API documentation generation ## 📋 Prerequisites - Python 3.10 or higher - AWS credentials configured (via `~/.aws/credentials`, environment variables, or IAM roles) - Docker (recommended) - [SyntropAIBox](https://test.pypi.org/project/syntropaibox/) core library ## 🐳 Docker Installation (Recommended) ### Build and Run ```bash # Build the image docker build -t mcp-server-aws-resources . # Run with AWS profile docker run -i --rm \ -e AWS_PROFILE=default \ -v ~/.aws:/root/.aws \ mcp-server-aws-resources:latest # Run with environment variables docker run -i --rm \ -e AWS_ACCESS_KEY_ID=your_key \ -e AWS_SECRET_ACCESS_KEY=your_secret \ -e AWS_DEFAULT_REGION=us-east-1 \ mcp-server-aws-resources:latest ``` ## ⚙️ Claude Desktop Integration Add to your `claude_config.json`: ```json { "mcpServers": { "aws-resources": { "command": "docker", "args": [ "run", "-i", "--rm", "-e", "AWS_PROFILE=default", "-v", "/Users/yourusername/.aws:/root/.aws", "mcp-server-aws-resources:latest" ] } } } ``` ## 🛡️ Security Features ### AST-Based Validation - Prevents malicious code injection - Whitelisted imports and functions - Controlled execution environment ### Safe Execution - Timeout protection (2-second default) - Isolated namespace - JSON-serialized responses ### Example Safe Query ```python # User provides this code snippet: import boto3 ec2 = session.client('ec2') result = ec2.describe_instances() ``` The system: 1. ✅ Validates AST syntax 2. ✅ Checks allowed imports (`boto3` approved) 3. ✅ Executes in sandbox with timeout 4. ✅ Returns JSON-serialized results ## 🔧 Usage Examples ### List EC2 Instances ```python import boto3 ec2 = session.client('ec2') result = ec2.describe_instances() ``` ### Create S3 Bucket ```python import boto3 s3 = session.client('s3') result = s3.create_bucket(Bucket='my-unique-bucket-name') ``` ### Lambda Functions ```python import boto3 lambda_client = session.client('lambda') result = lambda_client.list_functions() ``` ## 🌟 SyntropAI Ecosystem Benefits ### Unified Multi-Cloud - Same patterns work across AWS, Azure, OCI - Consistent authentication and error handling - Provider-agnostic abstractions ### Non-Hardcoded Services - Supports **any** AWS service automatically - No service catalog limitations - Future services work immediately ### Enterprise Ready - Security-first design - Docker containerization - Comprehensive logging ## 🔗 Related Projects - **[Main Documentation](https://github.com/paihari/documentation-syntropai)**: Complete ecosystem overview and architecture - **[SyntropAIBox Core](https://test.pypi.org/project/syntropaibox/)**: Shared abstraction library - **[Azure MCP Server](../mcp-server-azure)**: Azure implementation - **[OCI MCP Server](../mcp-server-oci)**: Oracle Cloud implementation - **[Finviz MCP Server](../mcp_finviz)**: Financial data server ## 🏆 Technical Highlights This implementation showcases: - **Advanced Abstraction Patterns**: Clean separation of concerns - **Security Engineering**: AST validation and sandboxed execution - **Cloud Architecture**: Scalable, maintainable multi-cloud design - **DevOps Excellence**: Containerized, configurable deployment ## 📞 Support For questions about the SyntropAI MCP ecosystem: - **Documentation**: [SyntropAI Documentation Project](https://github.com/paihari/documentation-syntropai) - **Author**: Hari Bantwal (hpai.bantwal@gmail.com) --- *This server demonstrates cutting-edge cloud abstraction technology, providing secure, unified access to AWS services through innovative architectural patterns.*