Awesome Agentic DevOps

A curated, official-first catalog of MCP servers, agent skills, AI agents, frameworks, and supporting resources for DevOps, Cloud, SRE, and Platform Engineering. It includes vendor and open-source project resources alongside a clearly labeled community-driven section.

Most agent lists stop at discovery. This one is built for operators:

• Official-first, community-inclusive — 80 entries organized into 15 catalog sections; official vendor and project resources are prioritized, while community-driven entries are separated in a dedicated community section.
• Scored, not just listed — every entry records action capability, human-approval controls, tracing evidence, maturity, and operational risk (how entries are scored).
• Audited by CI — GitHub repository entries are checked weekly for reachability, archived status, and stale push activity; non-GitHub documentation links are outside this automated check and require curator review.
• Installable, not just readable — one command installs hundreds of skills from cataloged Google, Microsoft, Azure, Azure DevOps, and Harness sources — plus a separate community set — into Claude Code, Cursor, Codex, VS Code, or Antigravity.
• Runnable, not just theoretical — documented reference-agent examples cover Terraform plan review, drift detection, and policy-driven container image releases.

Safety first: some cataloged tools and agents can change infrastructure. Prefer read-only or proposal mode, require human approval before write actions, and use least-privilege credentials — full guidance in the safety model and the practical operator safety checklist.

Contents

• Evaluation labels
• Compliance evidence checklist
• Operator safety checklist
• Top picks by use case
• Install skills into your coding agent
• Recently added
• Local reference agents
• Curated catalog
• How to contribute

Evaluation labels

Labels are shorthand for structured fields recorded on every entry in data/repos.yaml — how entries are scored explains each field and how it is verified.

Compliance evidence checklist

Production-adjacent agent runs need a reviewable evidence packet, not just a chat transcript. Use the compliance evidence checklist to capture request context, identity and data boundaries, redacted tool calls, approval records, validation output, and follow-ups for MCP servers, skills, incident copilots, Terraform reviewers, and other DevOps agents.

At minimum, record an explicit approval before any write-capable agent mutates infrastructure, source control, CI/CD, identity, secrets, incidents, or production telemetry configuration. If the approval cannot be captured, keep the agent in read-only or proposal mode.

Operator safety checklist

Before wiring any catalog entry into cloud accounts, clusters, CI/CD, identity, secrets, databases, observability, or incident-response systems, run through the operator safety checklist. It gives teams a concrete preflight for read-only-first evaluation, domain-specific credential boundaries, no-secret-in-context handling, dry-run/proposal evidence, approval gates, blast-radius limits, and audit artifacts.

Top picks by use case

Install skills into your coding agent

Install hundreds of Agent Skills from the cataloged Google, Microsoft, Azure, Azure DevOps, and Harness sources with one command:

# Claude Code, macOS/Linux — installs every official skill into ~/.claude/skills/
curl -fsSL https://raw.githubusercontent.com/DevOpsAIguru123/awesome-agentic-devops/main/install/claude-code/install.sh | sh -s -- --official

Choose breadth with a flag: --official (361 skills), --community (26), or --all (387). Python 3 is the only requirement — there is no pip install step.

Pass --dry-run to preview first. Commands for Cursor, Codex, VS Code, Antigravity (Mac and Windows) — plus target folders and safety notes — are in the install guide. To install just one company or product, use the official skills catalog or the community skills catalog.

Recently added

New catalog entries appear below; notable non-entry changes (formatting, labels, tooling) are in the changelog.

Local reference agents

This repo keeps runnable reference agents under agents/.

Curated catalog

The source of truth is data/repos.yaml. The catalog combines official vendor and open-source project resources with a dedicated section for community-driven tools and references. Entries include executable software as well as SDKs, documentation, security frameworks, registries, and other ecosystem resources.

Official Cloud MCP Servers and Agent Toolkits

Official DevOps MCP Servers

Official Security and Code-Quality MCP Servers

Official CI/CD and GitOps MCP Servers

Official MCP SDKs, Reference Implementations, Registries, and Governance Platforms

Official CloudOps Agent Samples

Official IaC MCP Servers

Official SRE MCP Servers

Official Agent Skills and Frameworks

Official Platform Agent Toolkits

Official Diagramming and Architecture MCP Tools

Official Browser Automation and Debugging MCP Servers

Official Data Platform MCP Servers

Official FinOps and Cloud-Cost MCP Servers

Community Discovery and Skills

How to contribute

Start with CONTRIBUTING.md. New entries should update data/repos.yaml, include a real operational use case, classify action level, and explain safety or approval behavior.

Run validation before opening a PR:

python scripts/validate_repos_yaml.py
python scripts/sync_readme_counts.py   # refreshes the intro's entry/category counts
pytest -q
python scripts/run_mock_eval_scenarios.py
python scripts/audit_github_repos.py --workers 12 --fail-on-unreachable