awesome-agentic-devops

Awesome Agentic DevOps is a curated catalog of AI agents and MCP servers for infrastructure work: cloud automation, DevOps pipelines, incident response, security, and infrastructure-as-code. Unlike general agent lists that stop at discovery, this one evaluates each entry by whether it is safe to run near production systems, whether it requires human approval before making changes, whether it preserves audit evidence, and how mature the project is.

The catalog organizes entries into eleven categories covering official cloud provider toolkits (AWS, Azure, Google Cloud), source-control platforms (GitHub, GitLab, Atlassian), CI/CD pipelines (Jenkins, ArgoCD), security and code quality tools (SonarQube, Okta, Snyk, Wiz), infrastructure-as-code integrations (Terraform, Pulumi), SRE and observability tools (Grafana, Datadog, PagerDuty), and agent frameworks. Each entry carries labels indicating whether it is production-adjacent, has approval gates, has tracing or audit output, and whether write actions require special review.

The project includes a top-picks table organized by use case so that an engineer evaluating AI automation for, say, AWS infrastructure or Terraform workflows can jump directly to the recommended starting point with a brief reason. The source of truth is a YAML file at data/repos.yaml that backs the readable catalog in the README.

A safety disclaimer appears prominently: agents in this catalog may touch real infrastructure, so the project recommends starting in read-only or proposal mode, requiring human approval before any write actions, and never placing secrets in model context.

The repository is aimed at DevOps and platform engineers evaluating AI automation, SREs designing incident-response copilots, security reviewers assessing infrastructure-agent risk, and developers building portfolio-quality reference agents. The README does not specify a license.