Breaking News: AWS MCP Server Now Generally Available

AWS today announced the general availability of the AWS MCP Server, a managed remote Model Context Protocol (MCP) server that provides AI agents and coding assistants with secure, authenticated access to all AWS services through a compact set of tools.

“This solves a critical challenge: giving an AI agent real access to AWS without handing over the keys to the kingdom,” said Jane Doe, AWS product manager for Agent Tooling. “Developers can now trust their agents to work with current documentation and best practices.”

Background: The Agent Access Problem

AI coding agents have proven useful for many tasks, but they struggle when working with AWS at any meaningful depth. Without access to current documentation, agents rely on potentially months-old training data, missing newer services like Amazon S3 Vectors or Amazon Bedrock AgentCore.

When asked to build infrastructure, agents often default to the AWS CLI instead of CDK or CloudFormation, and they generate IAM policies that are far broader than necessary. The result: infrastructure that works in a demo but isn’t production-ready.

How the AWS MCP Server Works

The AWS MCP Server addresses this with a small, fixed set of tools that do not consume the model’s context window. The call_aws tool executes any of over 15,000 AWS API operations using existing IAM credentials, with new APIs supported within days of launch.

The search_documentation and read_documentation tools retrieve current AWS documentation and best practices at query time, ensuring the agent always works from up-to-date information. “This eliminates the dependency on stale training data,” Doe explained.

New Capabilities in GA Release

With general availability, several new capabilities debut. The AWS MCP Server now supports IAM context keys, allowing fine-grained access without a separate IAM permission for the server itself. Documentation retrieval no longer requires authentication, and token consumption per interaction has been reduced for complex multi-step workflows.

The new run_script tool lets agents write short Python scripts that run server-side in a sandboxed environment. The sandbox inherits IAM permissions but has no network access, enabling data processing without exposing the local file system or shell. “Agents can chain API calls, filter responses, and compute results in a single round-trip,” said Doe. “That’s faster and more context-efficient.”

From SOPs to Skills

The most significant addition is the transition from Agent SOPs to Skills. Skills provide curated guidance and best practices for specific tasks, helping agents build more effectively on AWS. This marks a shift from static procedures to dynamic, updateable knowledge.

What This Means for Developers

The AWS MCP Server enables developers to delegate complex AWS tasks to AI agents with confidence. Secure, authenticated access means agents can now produce production-ready infrastructure using CDK or CloudFormation with appropriately scoped IAM policies.

“For the first time, agents can build infrastructure that you’d be willing to deploy in production,” Doe emphasized. The combination of real-time documentation retrieval and sandboxed script execution positions the MCP Server as a key component in the Agent Toolkit for AWS.

As AWS continues to launch new APIs, the MCP Server will support them within days, ensuring agents remain current. This GA release marks a significant step toward reliable, secure AI-assisted cloud development.