NVIDIA-NeMoOfficial·12 Agent Skills Included

Nemotron

Train, fine-tune, and deploy NVIDIA Nemotron open AI models

Guides AI agents through the full Nemotron model lifecycle: data curation, synthetic data generation, fine-tuning, RL alignment, evaluation, and deployment. Eliminates guesswork when chaining training steps, picking GPU configurations, or writing safety policies for guardrail models. Includes ready-made runbooks for Text2SQL LoRA fine-tuning, retrieval recipes, airgapped cluster deployment, and model knowledge bases.

2,015|411

↗

npx skills add NVIDIA-NeMo/Nemotron --all -g -y

All Skills in This Repository (12)

Pure Emerald Level Indicators

📦 In Repo

NVIDIA-NeMo

nemotron-customizer-airgap

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

nemotron-customize

Plans NVIDIA Nemotron customization pipelines and generates minimal YAML/config/run commands from step manifests.

nemotron-add-pattern

Generate Nemotron step pattern files with frontmatter and validate via pytest.

nemotron-super3

Retrieve cited facts about NVIDIA Nemotron 3 Super architecture and training pipeline from indexed references.

nemotron-add-step

Add a Nemotron step to the catalog with manifests and wiring.

nemotron-add-model

Create standardized Nemotron model knowledge bases with YAML frontmatter and validated marketplace registration.

nemotron-nano3

Retrieve repo-cited facts about Nemotron 3 Nano architecture and training.

nemotron-3-5-lightning-text2sql-lora

Runs LoRA fine-tuning of Nemotron-3.5 Lightning on BIRD Text2SQL data with Megatron-Bridge.

nemotron-3-ultra-text2sql-lora

Runs Nemotron-3 Ultra Text2SQL LoRA fine-tuning as multi-node SLURM jobs via Megatron-Bridge.

nemotron-policy-generator

Generates custom content-safety policies for NVIDIA Nemotron guardrail models as Markdown, JSON, and inference prompts.

nemotron-ultra

Answers factual questions about NVIDIA Nemotron 3 Ultra architecture, training, and deployment.

nemotron-retrieval-recipes

Plan, debug, tune, evaluate, and deploy Nemotron embed and rerank retrieval recipes.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install Nemotron skills?▼

Run `npx skills add NVIDIA-NeMo/Nemotron --all -g -y` in your terminal to install all skills in this suite globally.

What can I do with the Nemotron skills?▼

You can plan and run model training pipelines, fine-tune models with LoRA, build content-safety policies, tune embedding and reranking models, and get cited answers about Nemotron model architecture and benchmarks.

How do I fine-tune a Nemotron model for Text2SQL?▼

The included runbook skills walk an agent through data prep, checkpoint conversion, LoRA training, and merging, automatically picking the right GPU configuration for your hardware.

Do these skills work with Claude Code and Codex?▼

Yes. All skills follow the standard SKILL.md format and were benchmarked with Claude Code and Codex agents.

Can I deploy Nemotron models in an airgapped cluster?▼

Yes. The airgap skill helps your agent build offline container image bundles and submit training steps inside clusters without internet access.