Language: 日本語English

nvidia Codex Plugin

Author
NVIDIA
Category
Developer Tools
Topics
AI Agents & AI App Development · Design & Media Production · Cloud, Deployment & CI/CD
Version
1.0.2
First seen
2026-07-12
Last confirmed
2026-07-13
Explanation last updated
2026-07-12

The explanation below is AI-generated. Please verify it against the sources.

nvidia is a plugin (version 1.0.2) authored by NVIDIA that bundles skills covering GPU acceleration, CUDA, AI agents, inference, robotics, Physical AI, Omniverse, and simulation. According to the plugin manifest, it is designed to help users understand the pieces of NVIDIA's ecosystem, choose a path, validate their setup, and build practical NVIDIA-powered workflows. It is categorized under Developer Tools with Interactive and Write capabilities, and is licensed under Apache-2.0 AND CC-BY-4.0. Per the README, this plugin is generated from plugins.d/nvidia.yml via a build script for delivery to the official OpenAI marketplace, and is distinct from the self-hosted NVIDIA/skills marketplace referenced on the homepage, which lists 229 skills across domains such as AI and Machine Learning, Physical AI, and Accelerated Computing.

Overview

According to the plugin manifest and homepage, this is a curated skill collection sourced from NVIDIA's official repository (NVIDIA/skills), which can also be installed independently for agents like Claude Code or Codex via 'npx skills add NVIDIA/skills'. It functions as a guided, interactive assistant across NVIDIA's technology stack rather than a standalone SaaS product.

What you can do with nvidia

  • Research NVIDIA CUDA-X libraries relevant to simulation and robotics, per the plugin's default prompt
  • Get guidance on turning real-world driving footage into a 3D neural scene for simulation
  • Get help deploying a secure OpenClaw agent
  • Use domain-specific skills such as rag-blueprint (deploy/configure/troubleshoot NVIDIA RAG Blueprint), cuopt-developer/cuopt-install (build and install NVIDIA cuOpt), deepstream-dev/deepstream-import-vision-model (DeepStream video-analytics pipelines), and omniverse-usd-performance-tuning/omniverse-cad-to-simready (Omniverse/USD workflows), as listed on the homepage
  • Browse and filter the homepage's 229 skills by domain (AI and Machine Learning, Physical AI, Accelerated Computing, Developer Tools, Infrastructure) or audience (Developer, AI Engineer, ML Engineer, and others)

Sources

Original description (English)

Skills for NVIDIAs ecosystem spans GPU acceleration, CUDA, AI agents, inference, robotics, Physical AI, Omniverse, and simulation. This plugin helps you understand the pieces, choose a path, validate your setup, and build practical NVIDIA-powered workflows.

History of nvidia

Back to list