English
Workflow collection · 2026-10-18

Cloud & DevOps MCP Servers and Skills

Start from the system you operate: a cloud account, a Kubernetes cluster, infrastructure configuration, or a delivery pipeline. Compare established cloud guides with new official servers and skills, then open the setup guide for your environment.

19 guides · 14 MCP / 5 Skills

These are actual directory guides grouped by operational task. An MCP connection supplies tools; a skill supplies workflow instructions. Review the individual guide for credentials, transport, and the operations it can expose.

Choose your starting point

Your taskStart hereHow to choose
Inspect an existing cloud accountProvider and platform serversChoose the vendor you already use, then begin with inventory and logs before requesting changes.
Diagnose a Kubernetes workloadGoogle Kubernetes Engine MCPUse cluster, event, and workload context together rather than guessing from an application error alone.
Plan infrastructure changesInfrastructure and platform guidanceUse Terraform context or provider skills to prepare a plan; review live resource operations separately.
Investigate a delivery failureDelivery and workflow operationsStart with pipeline or flow-run evidence, then follow the affected service into its provider guide.

Cloud platforms and Kubernetes

Choose the integration matching the account and platform you actually operate.

Infrastructure and platform guidance

Connect infrastructure context with instructions for configuring and operating your platform.

Delivery and workflow operations

Follow build, container, deployment, and background-job context into the next diagnostic step.

Questions about this collection

Should I start with an MCP server or a skill?

Use a server when the task needs connected resource context or tools. Use a skill when the assistant needs platform-specific instructions. Some plugins bundle both; each guide explains its installation path.

Do these guides deploy anything automatically?

Reading a directory guide does not connect an account or run a tool. After installing a server yourself, its credentials and exposed tools determine the operations your client can request.

What should I open after investigating a failed deployment?

Follow the cloud provider guide for resource state and the observability collection for telemetry. Keep the same application, environment, and incident time window when moving between tools.

Keep exploring

Explore the fourth batch: 29 official cloud, payment, security, and monitoring additions →

See all 25 additions in the third batch → · Full directory →