Language: 日本語English

aws-agents Plugin

Author
Amazon Web Services
Category
Development
Topics
AI Agents & AI App Development · Cloud, Deployment & CI/CD
First seen
2026-07-09
Last confirmed
2026-07-13
Explanation last updated
2026-07-12

The explanation below is AI-generated. Please verify it against the sources.

aws-agents is a plugin within AWS's official Agent Toolkit for AWS that helps AI coding agents build, deploy, and operate AI agents on AWS, using Amazon Bedrock AgentCore as the primary runtime. According to the aws-agents README, it provides 7 skills covering the full agent lifecycle, from scaffolding a new project to production hardening, using progressive disclosure to load reference material only when needed. It bundles an MCP server called awsknowledge for AWS documentation and service reference, and can be installed in Claude Code or Codex via plugin marketplace commands. Prerequisites listed in the README include AgentCore CLI v0.9.0+, AWS CLI with configured credentials, Node.js 20+, and Python 3.11+ with uv. The plugin is licensed under Apache-2.0.

Overview

aws-agents is a plugin distributed as part of AWS's Agent Toolkit for AWS (a GitHub repository of AWS-supported MCP servers, skills, and plugins). It bundles agent skills and an MCP server configuration for coding agents such as Claude Code, Codex, and Cursor, focused specifically on building and running AI agents on AWS with Amazon Bedrock AgentCore.

What you can do with aws-agents

  • Scaffold a new AI agent project on AWS (skill: agents-get-started)
  • Add capabilities such as memory, VPC networking, multi-agent setups, or migration from Bedrock (skill: agents-build)
  • Connect tools/APIs, add a gateway, or restrict tool access with Cedar policies (skill: agents-connect)
  • Deploy an agent, including handling CDK errors, rollback, and canary releases (skill: agents-deploy)
  • Debug environment/CLI issues and check setup via logs (skill: agents-debug)
  • Evaluate agent quality, observability/traces, and cost (skill: agents-optimize)
  • Run through a production checklist and security hardening before launch (skill: agents-harden)
  • Access AWS documentation and service reference through the bundled awsknowledge MCP server
  • Install the plugin into Claude Code or Codex via marketplace commands described in the README

Sources

Original description (English)

Build, deploy, and operate AI agents on AWS. Skills for scaffolding agents with Amazon Bedrock AgentCore, connecting tools, memory, policies, evaluation, debugging, and production hardening.

History of aws-agents

Back to list