AWS DevOps Agent Skills

Open-source skills for AWS DevOps Agent that extend its capabilities for incident response, root cause analysis, and operational troubleshooting.

Each skill provides domain-specific knowledge, decision trees, and step-by-step runbooks that the agent follows during investigations. Use them as-is to enhance your agent, or as templates for writing your own custom skills.

All skills were tested using Agent Skill Eval and manually in DevOps Agent web app, for functionality without skill and with skill, and for effective triggering. The tests reports are available in each skill's evals/ directory.

Note: Agent Skill Eval currently uses Claude CLI for functional, trigger, compare and report tests. We're researching ways for local automated testing using AWS DevOps Agent APIs. We complement the current Agent Skill Eval automated testing with manual tests using DevOps Agent web app.

Note: If you intend to deploy these skills in production, start with a non-production environment first.

What Are AWS DevOps Agent Skills?

AWS DevOps Agent skills are structured instruction sets that teach the agent how to investigate specific operational scenarios. Skills follow the open Agent Skills specification and can be uploaded to your DevOps Agent deployment to extend its knowledge beyond built-in capabilities.

Skills enable DevOps Agent to:

• Specialize with investigation procedures, best practices, and organizational knowledge specific to your infrastructure
• Automatically load relevant instructions during investigations, eliminating repetitive guidance
• Compose multiple skills for end-to-end investigation workflows (e.g., retrieving deployments from your CI/CD pipeline and searching code repositories)
• Guide the agent in using your custom MCP server tools effectively for infrastructure-specific workflows

Available Skills

Skill	Description	Agent Types	Author	Docs
aws-health-events	Retrieves and analyzes AWS Health events (service issues, scheduled changes, account notifications) to identify AWS-side events that correlate with operational issues	Chat tasks, Incident RCA	udid-aws	README
support-cases	Searches and analyzes AWS Support cases to find historical incidents with similar symptoms, proven remediations, and recurring patterns	Chat tasks, Incident RCA	udid-aws	README
eks-operation-review	Performs comprehensive Amazon EKS operational reviews aligned with the AWS EKS Best Practices Guide covering security, reliability, networking, and scalability	Chat tasks, Prevention	yakiratz-aws	README
rds-operation-review	Performs comprehensive Amazon RDS operational reviews	Chat tasks, Prevention	yakiratz-aws	README

Getting Started

1. Clone the repository

git clone https://github.com/aws-samples/sample-code-for-devops-agent-skills.git
cd sample-code-for-devops-agent-skills

2. Choose a skill

Browse the skills table above and read the skill's README.md for details on its purpose, prerequisites, and sample prompts.

3. Upload to AWS DevOps Agent

Zip the skill directory (see the zip command in each skill's README) and upload it via the AWS DevOps Agent Operator Web App. Detailed instructions are in each skill's README.

Skill Directory Structure

Each skill follows a consistent structure based on the Agent Skills specification:

skills/<skill-name>/
├── SKILL.md          # Main skill instructions with frontmatter (required)
├── README.md         # Documentation, prerequisites, and upload guide
├── CHANGELOG.md      # Version history
├── evals/            # Evaluation queries and benchmarks
├── assets/           # Images, diagrams, data files (optional)
└── references/       # Supplementary reference docs (optional)

The SKILL.md, references/, and assets/ directories are what AWS DevOps Agent reads at runtime. Everything else supports development, testing, and documentation.

Writing Your Own Skills

Want to create custom skills for your operational workflows? See the AWS DevOps Agent skills documentation for the full guide, or use the skills in this repository as templates.

Key principles for effective skills (see also the Agent Skills best practices):

• Decide which agent types in DevOps Agent are relevant for your skill
• Write a description that specifies when and why the skill should activate — include specific symptoms, services, or error patterns that trigger it
• Ground instructions in real expertise — specific API patterns, edge cases, and project conventions, not generic advice
• Keep SKILL.md focused and under 500 lines; move detailed reference material to references/
• Add what the agent wouldn't know on its own — omit explanations of general concepts
• Favor step-by-step procedures over declarative statements so the approach generalizes across tasks
• Include decision trees for branching scenarios and checklists for multi-step workflows
• Provide defaults rather than menus — pick a recommended approach and mention alternatives briefly
• Include a gotchas section for non-obvious facts that defy reasonable assumptions
• Test with the Agent Skill Eval framework, and manually using the DevOps Agent web app, without skill and with skill

Contributing

We welcome contributions of new skills and improvements to existing ones. See CONTRIBUTING for guidelines.

References

AWS Documentation

• AWS DevOps Agent product page
• AWS DevOps Agent User Guide
• AWS DevOps Agent API Reference
• AWS DevOps Agent Skills — Creating and uploading skills

Blog Posts and Articles

• Extend AWS DevOps Agent with Custom Skills for Your Operational Workflows
• Building an End-to-End Agentic SRE Using AWS DevOps Agent
• Best Practices for Deploying AWS DevOps Agent in Production
• Leverage Agentic AI for Autonomous Incident Response with AWS DevOps Agent

Specifications and Tools

• Agent Skills specification — the open standard this project follows
• Agent Skill Eval — evaluation framework for testing skills

License

This library is licensed under the MIT-0 License. See the LICENSE file.