gke-golden-path
Provides GKE golden path configuration defaults, production readiness checklists, and cluster default patterns. Use when designing GKE clusters, verifying GKE production readiness, or checking configurations against GKE defaults. Don't use for setting up node autoscaling specifically (use gke-scaling instead).
Security Assessment
About gke-golden-path
gke-golden-path is a configuration-reference skill that codifies Google's recommended "golden path" defaults for production Google Kubernetes Engine (GKE) Autopilot clusters. It solves the problem of scattered, hard-to-remember best practices by providing a single opinionated set of sensible cluster defaults, production-readiness checklists, and guidance on which decisions are effectively irreversible once a cluster is created.
The skill distinguishes "always-apply" defaults (such as private nodes, Workload Identity, Secret Manager rotation, Shielded VM secure boot, advanced datapath, vertical pod autoscaling, restricted Pod Security Standards, and a broad monitoring/logging component set) from customer-configurable settings (like DNS external traffic, max pods per node, machine and boot-disk types, and maintenance exclusion windows) that may be changed with justification. It emphasizes flagging Day-0 versus Day-1 choices, documenting rationale for deviations, and a tool preference order of MCP tools (get_cluster, create_cluster, update_cluster) over gcloud over kubectl. Guardrails explicitly forbid requesting or outputting secrets and encourage discovering project context through tooling rather than asking users to paste IDs.
It targets platform engineers, SREs, and DevOps practitioners designing new GKE clusters, auditing existing clusters against a known-good baseline, or verifying production readiness. Because it is a defaults-and-checklist reference (with an accompanying golden-path-autopilot.yaml asset) rather than an executor, it is most useful as advisory guidance during cluster design and configuration review.
FAQ
What does this skill provide?
It provides GKE Autopilot golden-path configuration defaults, a split between always-apply best practices and customer-configurable settings, production-readiness guardrails, and a reference YAML of full cluster-level policy settings.
When should I not use it?
The skill explicitly directs you to the separate gke-scaling skill for node autoscaling, and it is not intended for database administration or workload-specific tuning beyond cluster defaults.
What are Day-0 versus Day-1 decisions?
Day-0 decisions like networking, private nodes, subnets, and IP allocation are hard or impossible to change after cluster creation, so the skill instructs asking clarifying questions before proceeding; Day-1 features are proposed as golden-path defaults with trade-offs for the customer to confirm.
Does it handle secrets?
No. Its guardrails explicitly prohibit requesting or outputting secrets such as tokens, keys, or service account JSON, and recommend discovering project context via MCP tools or gcloud config rather than asking users to paste sensitive identifiers.
Which tools does it prefer for applying changes?
It prefers MCP tools first, then gcloud, then kubectl, but respects a user's explicit request to use gcloud or kubectl for the session.
Install gke-golden-path
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
google/skills