skills.rest
Lhhiep-maxcode
Lhhiep-maxcodeCOMMUNITY
Explore all skills by this author →
GitHub

nemotron-customizer-airgap

Validate Nemotron Customizer steps and build airgap image bundles for offline deployment.

Updated Jul 30, 2026
Search Tags:#docker#nemotron#workflow validation#nvidia#container images#offline deployment#airgap
One-click install
npx skills add https://github.com/Lhhiep-maxcode/Nemotron --skill nemotron-customizer-airgap-lhhiep-maxcode
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: nemotron-customizer-airgap
Source: https://github.com/Lhhiep-maxcode/Nemotron/tree/main/deploy/nemotron-customizer/airgap
Command: npx skills add https://github.com/Lhhiep-maxcode/Nemotron --skill nemotron-customizer-airgap-lhhiep-maxcode

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill helps teams prepare Nemotron Customizer workflows for airgapped environments, removing the uncertainty of which images, dependencies, and runtime overlays are needed before offline deployment.

Core Features & Use Cases

  • Airgap Planning: Select workflow stages, expand dependencies, and validate required step files and configs before building.
  • Image Assembly: Group selected steps into execution images, discover missing Python packages, and build launcher and execution containers on a connected machine.
  • Offline Submission: Save tarballs, review manifests, and run Nemotron steps in an airgapped cluster using the baked images and persistent storage mounts.
  • Use Case: An operator wants to run SFT Megatron-Bridge offline, so the skill plans the bundle, bakes repo overlays into the execution image, and generates the runtime overlay config to prevent GitHub cloning.

Quick Start

Ask the AI to plan and prepare an airgap bundle for the selected Nemotron Customizer steps, then report the required execution images, saved tarballs, and offline runtime settings.

Frequently Asked Questions about nemotron-customizer-airgap

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

FAQPage Schema
How do I prepare Nemotron Customizer workflows for offline deployment?▼

To prepare Nemotron Customizer workflows for offline deployment, you validate selected steps, expand dependencies, and build airgap image bundles. This process prevents runtime cloning by baking repository overlays directly into execution images.

What is an airgap image bundle and when do I need it for Nemotron?▼

An airgap image bundle is a packaged set of Docker containers and manifests needed to run workflows in an offline cluster. You need it when deploying Nemotron steps in environments without internet access to avoid cloning repositories during runtime.

How do I build Docker container images for an airgapped cluster?▼

You build Docker container images for an airgapped cluster by grouping selected workflow steps into execution images on a connected machine, discovering missing Python packages, and saving the final containers as tarballs for offline submission.

Does the Nemotron airgap bundler support resuming interrupted image builds?▼

Yes, the bundler supports resume-safe image build planning for selected Nemotron step targets. It tracks the build state to allow recovery from interruptions without restarting the entire execution image assembly process.

How do I configure runtime overlays to prevent GitHub cloning in offline clusters?▼

You configure runtime overlays by generating persistent-storage-based runtime configuration and baking repository overlays into the execution image. This ensures the airgapped cluster uses local files instead of attempting GitHub cloning.

What are the limitations of running SFT Megatron-Bridge offline?▼

Running SFT Megatron-Bridge offline requires completing step catalog validation and execution-image grouping on a connected machine first. You must also ensure persistent storage mounts are correctly configured for the airgapped cluster submission.

Related Skills

Explore More Software Engineering Skills→
NVIDIA-NeMoNVIDIA-NeMo
1.8k

nemotron-customizer-airgap

Plan and build Nemotron Customizer airgap image bundles from airgap.yaml.

Official
Advanced
NVIDIANVIDIA
14

k8s-launch-kit-dryrun

Preview and validate Kubernetes deployment manifests generated by k8s-launch-kit without applying cluster changes.

Official
Intermediate
hieutrtrhieutrtr
8

docker-best-practices

Standardize Docker containerization for Python and React projects with multi-stage builds.

Community
Advanced
logicalclockslogicalclocks
15

hops-online-inference

Deploys ML models and runs online inference in Hopsworks via Python SDK.

Official
Advanced