AWS MCP Servers Skill
Official AWS MCP servers as agent skills — docs, pricing, CDK, and service APIs for AI agents.
About
This is the official AWS Labs collection of Model Context Protocol servers, packaged as agent skills, giving AI coding agents structured access to AWS documentation, service APIs, billing and cost data, CDK/IaC guidance, and infrastructure metadata. Maintained by AWS under Apache-2.0 with frequent updates tracking service changes, it includes servers for Lambda, S3, DynamoDB, CloudWatch, Cost Explorer, and more, distributed as Docker images via the public AWS ECR gallery.
Agents use your existing AWS credentials (CLI profiles/IAM roles) — no separate token to manage. The honest caveats: it's MCP servers rather than SKILL.md prompt packs, so setup involves Docker or uv; and like any cloud-automation tooling, agents with write access need guardrails before touching production infrastructure.
What it can do
- AWS documentation MCP server
- Service API access (Lambda, S3, DynamoDB, more)
- Cost Explorer and pricing data
- CDK and IaC best-practices guidance
- CloudWatch observability access
- Apache-2.0, AWS-maintained
Setup
Pull the Docker images from the public AWS ECR gallery (gallery.ecr.aws/awslabs-mcp) or build from https://github.com/awslabs/mcp. Configure your agent's MCP client with the server entries and your AWS credentials.
Prerequisites
AWS credentials (CLI profile or IAM role); Docker or uv for running the servers.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 49; Trending score 80/100 from researcher assessment.
★ 9,700 GitHub stars
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.