NVIDIA-OmniverseOfficial·35 Agent Skills Included

ovrtx

RTX sensor simulation, rendering, and USD scene control

Builds camera, lidar, and radar sensor simulation into Python and C/C++ applications using NVIDIA Omniverse RTX. Covers renderer setup, USD scene loading, attribute reads and writes, pointcloud decoding, picking, and render output mapping. Removes guesswork with tested code snippets, migration guides, and ready-made patterns for async operations and GPU interop.

201|25

↗

npx skills add NVIDIA-Omniverse/ovrtx --all -g -y

Available:

Routes each user request to the correct task skill, explains the ovrtx and ovstage ownership split, and defines the GPU, driver, and runtime checks required before claiming validation.

All Skills in This Repository (35)

Pure Emerald Level Indicators

📦 In Repo

NVIDIA-Omniverse

adding-doc-snippets

Automate creation and validation of tested code snippets for documentation pages.

combining-scene-with-runtime-prims

Combine a base USD scene with inline runtime prims via subLayers composition.

status-queries

Query asynchronous ovrtx operation status via Python and C APIs.

error-handling

Handle Python and C API errors in ovrtx applications.

configuring-lidar-sensors

Automate OmniLidar prim configuration and PointCloud channel mapping in USD scenes.

writing-attributes

Write scalar and array USD prim attributes with dtype and shape validation.

camera-outputs-rt2

Identify available RT2 camera outputs and recommended readback methods.

reading-render-output

Map Ovrtx render outputs to accessible pixel data on CPU or CUDA.

writing-transforms

Write 4x4 transform matrices to USD prims using Python or C/C++.

interpreting-radar-pointclouds

Interpret radar PointCloud tensors by mapping channel meanings and validating detections.

project-setup-c

Scaffold a CMake C/C++ project with ovrtx dependencies.

attribute-bindings

Implement persistent attribute bindings for repeated OVRTX prim writes.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install ovrtx?▼

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

What does ovrtx do?▼

ovrtx embeds Omniverse RTX rendering and physically accurate camera, lidar, and radar simulation into Python or C/C++ applications. It targets robotics learning, synthetic data generation, and industrial simulation workflows.

How do I simulate lidar or radar sensors?▼

The suite includes dedicated skills for authoring OmniLidar and OmniRadar prims, requesting PointCloud output channels, and reading or interpreting the resulting detection tensors.

What hardware do I need to run ovrtx?▼

You need an NVIDIA RTX-capable GPU with a supported driver, plus unsandboxed execution and internet access for examples that load remote assets.

How do I migrate from ovrtx 0.3 to 0.4?▼

Dedicated migration skills move your Python or C/C++ code from renderer-owned scene APIs to the new ovstage-attached model, with step-by-step API mapping tables.