Skills from NVIDIA’s repositories that passed our trust, licence and security gates. Each keeps its author’s attribution.
| Repository | Tier | Verdict | Authorship | Checked | Skills |
|---|---|---|---|---|---|
| NVIDIA/skills | Vendor | Approved · created 2026-02-25, pushed 2026-11-06, 3888 stars | Written in-house | 2026-11-07 | 431 |
Files are mirrored only where the licence allows redistribution; otherwise we link to the source.
NVIDIAOfficial
Execute and evaluate the FoundationPose perception pipeline on BOP-format datasets using TAO depth engines.
NVIDIAOfficial
Train and deploy NVPanoptix3D models for panoptic 3D scene reconstruction and occupancy completion.
NVIDIAOfficial
Numerical optimization API for solving LP, MILP, and QP problems using NVIDIA cuOpt.
NVIDIAOfficial
Perform KPI analysis for object detection models by comparing inference results against ground truth.
NVIDIAOfficial
Add a new cuTile GPU kernel operator to TileGym. Covers dispatch registration in ops.py, cuTile backend implementation, __init__.py exports, test creation, and benchmark in tests/benchmark. Use when adding, creating, or implementing a new cuTile operator/kernel in TileGym, or when asking how to register a new cuTile op.
NVIDIAOfficial
Upgrades a grover_base checkpoint to a hybrid cMIM-vocab model and continues pretraining on a custom corpus.
NVIDIAOfficial
Bridge socket-oriented host applications to a BlueField DPU peer without code changes using the DOCA Socket Relay.
NVIDIAOfficial
Provides BMC-driven day-1 platform bring-up workflows for NVIDIA BlueField-4 DPUs.
NVIDIAOfficial
Generates validated CDF simulation jobs for PCBA inspection from board USD files.
NVIDIAOfficial
Fine-tunes NV-Segment-CT VISTA3D models on CT NIfTI datasets using standard or fixed-channel softmax workflows.
NVIDIAOfficial
Perform a point-in-time health assessment of a Run:ai cluster, identifying node pool status, GPU availability, and scheduling blockers.
NVIDIAOfficial
Train, evaluate, and deploy lightweight SegFormer transformer models for real-time semantic segmentation tasks.
NVIDIAOfficial
A structured workflow for porting quantum circuits from frameworks like Qiskit into CUDA-Q kernels while maintaining algorithmic integrity.
NVIDIAOfficial
Standardized procedures for detecting, validating, and diagnosing DOCA software version consistency and build-vs-runtime drift.
NVIDIAOfficial
A systematic approach to selecting and verifying physics engines for Physical AI and robotics simulation workflows.
NVIDIAOfficial
Automated QA test runner for Holoscan Sensor Bridge hardware.
NVIDIAOfficial
Operational guide for configuring and sizing parallelism strategies in Megatron Bridge for dense and MoE models.
NVIDIAOfficial
Provides guidance for using doca_flow_tune to analyze, visualize, and optimize DOCA Flow pipelines.
NVIDIAOfficial
Edit and persist changes to existing Isaac for Healthcare workflow scenes and task contracts.
NVIDIAOfficial
Use for NVIDIA DGX Spark and related GB10 system identification, capacity, troubleshooting, known issues, updates, recovery, playbooks, examples, NVIDIA Sync, clustering, Custom Apps, Field Diagnostics, or support evidence. Always route through this skill. Do not use for Apache Spark or unrelated meanings of spark.
NVIDIAOfficial
Operate and debug the DOCA Firefly Service container for PTP-grade time synchronization on NVIDIA BlueField DPU.
NVIDIAOfficial
Retrieve and summarize version-specific official NVSHMEM and NVSHMEM4Py documentation.
NVIDIAOfficial
Fine-tune HuggingFace models on local NVIDIA GPUs using a standardized, reproducible NGC-based workflow.
NVIDIAOfficial
Integrate TileGym kernels into Hugging Face `transformers` models by replacing the library's submodule(s) and certain class(es)' implementations, and patching certain class(es)' init/forward/load weight methods prior to instantiating models. Used when the user requires integrating TileGym kernels into `transformers` models.