GKE Basics — Production-Ready Cluster Guide

An expert guide skill that plans, creates, and configures Google Kubernetes Engine (GKE) clusters using a golden-path Autopilot approach. Given a project/region/cluster name, it applies the right production defaults (private nodes, Workload Identity, Secret Manager, advanced datapath, VPC-native networking, vertical/horizontal autoscaling) and produces runnable gcloud + kubectl commands.

When to use it: when provisioning a new GKE cluster, designing network topology (private cluster, VPC, Gateway API), hardening security (RBAC, Binary Auth, gVisor), setting up autoscaling/ComputeClass, planning cost optimization (Spot VM, CUD, rightsizing), placing AI/ML inference (GPU/TPU, vLLM) workloads, upgrading a cluster, or setting up observability.

Output: a plain-English, source-referenced Markdown guide with a decision-summary table + Day-0/Day-1 checklists + runnable command blocks + a cost note. It clearly flags Day-0 (irreversible) decisions. All configurations are enriched with special warnings about network and IP decisions that cannot be changed after the cluster is created.