Language:JapaneseEnglish

aws-agents-for-devsecops Claude Code Plugin

Author
Amazon Web Services
Category
Development
Topics
Security, Auth & Compliance · Cloud, Deployment & CI/CD · Code Quality, Review & Testing
First cataloged
2026-07-09 (UTC)
Explanation last updated
2026-08-07 (UTC)
Source (GitHub) last updated
2026-08-06 (UTC) (21 days ago)

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

This is a Claude Code plugin that connects to two remote AWS services, AWS DevOps Agent and AWS Security Agent, through an MCP server. It ships as a set of skills, slash commands, and a bundled MCP server configuration rather than a standalone application. Access requires AWS SigV4 credentials or an access token for the target AWS account. The README also warns that responses from these agents may contain suggested commands or code that should never be executed automatically without explicit user approval.

About the service

AWS DevOps Agent and AWS Security Agent are remote, agent-based services operated by AWS. The plugin communicates with them through a bundled MCP server, exposing tool categories such as chat, incident investigation, release testing, release readiness review, agent space management, access token management, service listings, and evaluation. Connecting requires AWS account credentials, either as SigV4 signing or a bearer access token.

What you can do with aws-agents-for-devsecops

  • Set up authentication for AWS DevOps Agent and AWS Security Agent via dedicated setup skills or commands
  • Investigate production incidents such as errors, out-of-memory events, or alarms with AWS DevOps Agent
  • Ask AWS DevOps Agent questions about cost, architecture, or topology through a chat interface
  • Run UI and API release tests and retrieve their reports
  • Request a release readiness review before merging a pull request
  • Coordinate work across multiple AgentSpaces using SigV4 auth
  • Run full or diff-only code security scans with AWS Security Agent
  • Run penetration tests against live endpoints
  • Build STRIDE-based threat models for design documents
  • Fetch, triage, and remediate security findings

Sources

Back to list