AWS Resources MCP Server

What is AWS Resources MCP Server?

A Model Context Protocol server that runs Python code via boto3 to query or modify any AWS resources. Designed for Python developers who want natural-language-driven AWS management, it runs as a Docker container and supports both read and write operations based on the IAM user’s permissions.

How to use AWS Resources MCP Server?

Install the server via Docker (pull the image or build locally) or through Smithery. Configure AWS credentials using environment variables (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION) or an AWS profile mounted from ~/.aws. In Claude Desktop, add the server to your MCP configuration with the Docker run command. Invoke the tool aws_resources_query_or_modify by providing a Python code snippet that uses boto3 and sets a result variable.

Key features of AWS Resources MCP Server?

• Executes boto3 code to query or modify AWS resources.
• Runs via Docker – no git clone needed.
• Python-based for stability and easy contributions.
• AST-based code analysis for execution safety.
• Restricted environment with limited imports and built‑ins.
• JSON serialization with proper AWS object handling.

Use cases of AWS Resources MCP Server?

• Query AWS resources (e.g., list S3 buckets, describe EC2 instances).
• Debug and resolve AWS permission errors (e.g., DynamoDB access denied).
• Retrieve CodePipeline deployment status programmatically.
• Perform management operations (create, update, delete) within IAM limits.
• Automate recurring AWS tasks using natural language prompts.

FAQ from AWS Resources MCP Server

What makes this different from other AWS MCP servers?

The README cites three advantages: runs directly from a Docker image (no local clone), uses Python/boto3 for better stability on Mac, and is easy for Python developers to contribute back to.

What are the runtime dependencies?

The server is containerized in Docker. Inside the container, it uses Python with boto3, and allows imports of boto3, operator, json, datetime, pytz, dateutil, re, and time. No additional system dependencies are needed beyond Docker.

Where are AWS credentials configured?

Credentials can be passed via environment variables (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, optional AWS_SESSION_TOKEN, AWS_DEFAULT_REGION) or by mounting a local ~/.aws directory and setting the AWS_PROFILE environment variable.

Is the tool restricted to read-only operations?

No. The README explicitly states that operations are not limited to Read