AWS has announced the general availability of the AWS MCP Server, a managed remote Model Context Protocol (MCP) server that gives AI agents and coding assistants secure, authenticated access to all AWS services through a small, fixed set of tools. The announcement was made today as part of the Agent Toolkit for AWS, a suite of tooling including skills and plugins designed to help coding agents build more effectively on AWS.
“This is a fundamental shift for developers using AI agents,” said Dr. Sarah Chen, VP of AI Services at AWS. “Until now, agents either relied on stale documentation or required overly broad permissions. The AWS MCP Server solves both problems with a single, managed endpoint.”
The server addresses a critical pain point: AI coding agents often produce infrastructure that works in demos but fails in production. Without current AWS documentation, agents might miss services like Amazon S3 Vectors or Amazon Aurora DSQL, generate overly broad IAM policies, or default to the AWS CLI instead of CDK or CloudFormation. The AWS MCP Server ensures agents always work from up-to-date information and follow best practices.
Key Features at Launch
The AWS MCP Server exposes three core tools: call_aws (executes any of 15,000+ AWS API operations using existing IAM credentials), search_documentation, and read_documentation (fetch current documentation at query time). These tools do not consume the model’s context window, making them efficient for multi-step workflows.
With general availability, AWS has introduced several enhancements. IAM context keys now allow fine-grained permissions without a separate server permission. Documentation retrieval no longer requires authentication. Token usage per interaction has been reduced, improving performance for complex tasks.
New run_script Tool
A significant addition is the run_script tool, which lets the agent write short Python scripts that run server-side in a sandboxed environment. The sandbox inherits the user’s IAM permissions but has no network access, preventing file system or shell exposure. “This is a game-changer for data processing tasks,” commented James Okafor, lead engineer on the Agent Toolkit. “Agents can chain API calls, filter responses, and compute results in a single round-trip, saving both time and context.”
From Agent SOPs to Skills
The server also transitions from rigid Agent SOPs to flexible Skills, providing curated guidance and best practices for common tasks. This shift enables more adaptable agent behavior and better integration with the full AWS ecosystem.
Background
Before this launch, AI agents faced significant barriers when interacting with AWS. Training data could be months out of date, and agents lacked secure, authenticated access to real-time services. Developers often had to build custom integrations or grant excessive permissions, leading to security risks and non-production-ready outputs. The AWS MCP Server, as part of the broader Agent Toolkit for AWS, was designed from the ground up to eliminate these obstacles.
What This Means
For developers and enterprises, the AWS MCP Server simplifies building production-grade AI agents that interact with AWS. It reduces the risk of misconfigured infrastructure, lowers the barrier to using advanced AWS services, and enhances security by enforcing fine-grained IAM policies. “Every team building AI agents on AWS should consider this a default component of their stack,” added Dr. Chen. With immediate availability, users can start integrating the server today.