k8s-capi

What This Skill Does

The k8s-capi skill provides a comprehensive interface for managing Kubernetes clusters through the Cluster API (CAPI) framework. By integrating with the kubectl-mcp-server, this skill enables users to provision, scale, monitor, and manage the lifecycle of entire Kubernetes clusters directly from the OpenClaw agent. It abstracts complex infrastructure operations into simple, actionable tools, covering everything from listing cluster machines to managing machine deployments and monitoring cluster health via health checks.

Installation

To install this skill, use the OpenClaw hub command: clawhub install openclaw/skills/skills/rohitg00/k8s-capi

Use Cases

This skill is designed for DevOps engineers, SREs, and platform teams managing multi-cluster Kubernetes environments. It is ideal for:

• Automating the lifecycle management of workload clusters across various providers (e.g., AWS, GCP, Azure).
• Scaling worker node pools dynamically based on workload demands using MachineDeployment scaling tools.
• Troubleshooting cluster health by inspecting machine status, health check results, and deployment states.
• Standardizing cluster creation through the use of ClusterClasses for repeatable infrastructure templates.
• Accessing remote clusters quickly by retrieving kubeconfig files programmatically.

Example Prompts

1. "List all clusters in the 'production' namespace and show me their current provisioning phase."
2. "Scale the worker node machine deployment 'production-cluster-md-0' to 10 replicas."
3. "Check the status of all machines in 'staging-cluster' and tell me if any are in a failed state."

Tips & Limitations

• Permissions: Ensure the agent's underlying context has the necessary RBAC permissions on the management cluster to list and modify CAPI objects.
• Namespaces: Always specify the correct namespace if your clusters are isolated. If omitted, the tool typically defaults to 'default', which may lead to 'not found' errors if your resources reside elsewhere.
• Asynchronous Operations: Remember that CAPI operations are often asynchronous. Scaling or creating a cluster may take time for the infrastructure provider to provision resources. Always check the phase status using capi_cluster_get_tool before assuming the operation is complete.
• Error Handling: When running kubectl apply for new clusters, ensure your YAML manifests include all mandatory provider-specific fields to avoid validation failures.