k8s-launch-kit-generate

Generate Kubernetes YAML manifests for NVIDIA networking deployments from cluster configuration.

14|5|Updated Nov 4, 2025
One-click install
npx skills add https://github.com/NVIDIA/k8s-launch-kit --skill k8s-launch-kit-generate
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: k8s-launch-kit-generate
Source: https://github.com/NVIDIA/k8s-launch-kit/tree/main/skills/k8s-launch-kit-generate
Command: npx skills add https://github.com/NVIDIA/k8s-launch-kit --skill k8s-launch-kit-generate

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill eliminates the complexity of selecting compatible NVIDIA Kubernetes networking profiles and manually assembling the corresponding deployment manifests.

Core Features & Use Cases

  • Profile Selection: Choose among SR-IOV, host-device, RDMA-shared, IPoIB, MacVLAN, and Spectrum-X networking profiles.
  • Manifest Generation: Create tailored Network Operator values, custom resources, IP pools, network policies, and example workloads from cluster configuration.
  • Hardware-Aware Defaults: Resolve fabric, deployment type, multirail, Spectrum-X multiplane mode, and plane count from discovered NIC capabilities.
  • Use Case: Generate deployment files for a heterogeneous GPU cluster, restrict output to selected hardware groups, and preview the resulting resources before applying them.

Quick Start

Ask the skill to generate NVIDIA Kubernetes networking manifests from your cluster configuration using the appropriate deployment profile and save them to an output directory.

Frequently Asked Questions about k8s-launch-kit-generate

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

FAQPage Schema
How do I generate Kubernetes manifests for NVIDIA networking deployments?

Generate Kubernetes manifests for NVIDIA networking by providing your cluster configuration to automatically create tailored Network Operator values, custom resources, and IP pools for your selected deployment profile.

Which networking profiles are supported for Kubernetes GPU clusters?

Supported networking profiles include SR-IOV, host-device, RDMA-shared, IPoIB, MacVLAN, and Spectrum-X, allowing you to apply the generated manifests to diverse GPU cluster infrastructure configurations.

Can I use Spectrum-X multiplane networking with the Network Operator?

Yes, you can generate manifests for Spectrum-X multiplane networking that satisfy strict topology requirements, resolving multiplane mode and plane count directly from discovered NIC capabilities.

How does hardware-aware profile selection work for RDMA network manifests?

Hardware-aware profile selection resolves fabric, deployment type, and multirail configurations by validating discovered NIC capabilities, ensuring generated RDMA network manifests match your physical hardware topology.

Can I restrict manifest generation to specific hardware groups in a heterogeneous GPU cluster?

Yes, you can restrict manifest output to selected source hardware groups, allowing you to preview and generate deployment files tailored to specific nodes within a heterogeneous GPU cluster.

What are the limitations of automatically applying generated network manifests to a cluster?

Manifest generation enforces safe separation, meaning you must manually apply the generated YAML files to your cluster afterwards to satisfy Network Operator version constraints and prevent accidental deployments.