skills.rest
primatrix
primatrixOFFICIAL
Explore all skills by this author →
GitHub

gke-tpu

Plan, render, and apply GKE TPU Job manifests on TPU v6e/v7x.

5|Updated Feb 4, 2026
Search Tags:#kubernetes#gke#workload management#tpu#nodepool management
One-click install
npx skills add https://github.com/primatrix/skills --skill gke-tpu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gke-tpu
Source: https://github.com/primatrix/skills/tree/main/plugins/gke-tpu/skills/gke-tpu
Command: npx skills add https://github.com/primatrix/skills --skill gke-tpu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires gcloud, kubectl, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a set of commands to manage GKE TPU workloads, TPU nodepools, TPU topology, reservations, or Kubernetes Jobs on TPU v6e/v7x, streamlining the process of deploying and managing TPU resources.

Core Features & Use Cases

  • GKE TPU Workload Management: Plan, render, and apply multi-doc Job manifests for GKE TPU workloads.
  • Nodepool Management: Plan and create TPU nodepools with specific topology and machine types.
  • Workload Rendering: Render one multi-doc Job manifest to a specified directory.
  • Cleanup: Delete workload plans and nodepool plans for destructive actions.
  • Use Case: If you need to deploy a TPU-based machine learning workload on GKE, this Skill can help you plan and create the necessary resources.

Quick Start

Run the plan-nodepool command with the appropriate configuration file to plan a TPU nodepool.

Frequently Asked Questions about gke-tpu

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy a machine learning workload on GKE TPU v6e or v7x?▼

To deploy a GKE TPU workload, you can plan, render, and apply multi-doc Job manifests specifically configured for TPU v6e or v7x architectures. This streamlines the process of deploying and managing TPU resources on Google Kubernetes Engine.

What is the best way to create a TPU nodepool with a specific topology in GKE?▼

Creating a TPU nodepool involves planning and provisioning nodes with specific topology and machine types tailored to your workload. You can execute the plan-nodepool command with your configuration to set up the TPU nodepool.

Do I need gcloud and kubectl to manage Kubernetes Jobs on TPU?▼

Yes, you need both gcloud and kubectl installed and configured to manage GKE TPU workloads. These dependencies are required to interact with Google Kubernetes Engine and the Kubernetes API for resource provisioning.

How do I clean up TPU workload plans and nodepool plans after execution?▼

To clean up TPU workload plans and nodepool plans, you can use the Skill's dedicated cleanup commands for destructive actions. This ensures that obsolete job manifests and nodepool configurations are safely deleted from your environment.

How does rendering a multi-doc Job manifest work for GKE TPU resources?▼

Rendering a multi-doc Job manifest generates the Kubernetes configuration files for your TPU workload and outputs them to a specified directory. This allows you to review the exact manifests before applying them to your cluster.

Related Skills

Explore More Software Engineering Skills→
ssmleossmleo

gke-upgrades

Automate GKE cluster upgrade planning and execution with gcloud and kubectl.

Community
Advanced
wangx7wangx7

gke-batch-hpc

Configure Kueue job queues and deploy batch workloads on GKE clusters.

Community
Intermediate
primatrixprimatrix

deploy-cluster

Automate SkyPilot-managed TPU cluster deployment on Google Kubernetes Engine.

Official
Advanced
primatrixprimatrix

apply-resource

Manage GKE TPU Nodepool resources via XPK commands.

Official
Intermediate