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
Configure and evaluate adaptive plugin behavior for NeMo Relay applications to optimize performance and request handling.
NVIDIAOfficial
Builds a native, single-GPU cuDF proof of concept for query engines, focusing on adapter integration and buffer ownership.
NVIDIAOfficial
Converts cuTile Python GPU kernels (@ct.kernel) to cuTile.jl Julia equivalents. Handles kernel syntax translation, 0-indexed to 1-indexed conversion, broadcasting differences, memory layout (row-major to column-major), type system mapping, and launch API differences. Use when converting, porting, or translating cuTile Python kernels to Julia cuTile.jl, or debugging/optimizing existing Julia cuTile translations.
NVIDIAOfficial
Prepare and validate object-detection false-negative data for AnomalyGenNext generation workflows.
NVIDIAOfficial
Manage Cosmos-Embed1 video-text embedding workflows for retrieval, search, and fine-tuning.
NVIDIAOfficial
Plan, install, and validate NVSHMEM and NVSHMEM4Py across various environments.
NVIDIAOfficial
Manage the lifecycle and inference requests for TAO model microservices across multiple compute platforms.
NVIDIAOfficial
Visual ChangeNet for binary image classification and segmentation in AOI defect detection. Use when training, evaluating, exporting, or running inference for PCB defect detection or visual inspection, comparing image pairs for PASS/NO_PASS classification, or producing change-segmentation masks. Trigger phrases include "train Visual ChangeNet", "ChangeNet classify", "ChangeNet segment", "AOI defect detection", "PCB inspection model".
NVIDIAOfficial
Hardware-accelerated DEFLATE and LZ4 compression offload for NVIDIA BlueField DPU and ConnectX NICs using the DOCA Compress library.
NVIDIAOfficial
Run streamed Unitree G1 cloud-navigation demos in Isaac Sim with guarded gantry release. Use for demo, destination, Mission Control, or cleanup requests.
NVIDIAOfficial
A workflow for building and verifying the Holoscan SDK 5.x from source using the repository run wrapper.
NVIDIAOfficial
Build DeepStream GStreamer pipelines interactively. Use when the user asks about pipelines for video/image inference, detection, tracking, or streaming — including natural phrases like 'pipeline to infer on image', 'run inference on video', 'detect objects in stream', 'save inference output', 'deepstream pipeline', 'gst-launch pipeline', 'process video with detection', 'build a pipeline', or any request involving GStreamer/DeepStream elements (nvinfer, nvstreammux, nvtracker, etc.).
NVIDIAOfficial
Standardized documentation and docstring conventions for the NeMo-RL project.
NVIDIAOfficial
A discovery skill for querying available TAO models, application workflows, and AutoML capabilities within the TAO Skill Bank.
NVIDIAOfficial
Define and validate NVIDIA PVA operator specifications, test matrices, and evidence for OpGen handoff.
NVIDIAOfficial
Configure and validate sequence packing for efficient LLM and VLM training in Megatron-Bridge.
NVIDIAOfficial
Automate the selection, validation, and deployment of NVIDIA Dynamo Kubernetes recipes.
NVIDIAOfficial
Run iterative improvement for NVIDIA TAO CLIP / SigLIP image-text retrieval on attribute-labelled data. Use when a request combines retrieval evaluation, weak-attribute or caption-pair mining, repeated retraining, and a stopping condition based on a retrieval KPI, validation plateau, or iteration budget; the customer need not know the DEFT or People Attribute Search (PAS) names. The self-contained workflow performs dataset preparation, zero-shot evaluation, attribute gap analysis, caption-space k-NN mining, history-aware selection, retraining, and re-evaluation. Treat `tao-deft-pas` as shorthand for this canonical `tao-run-deft-pas` workflow. Do not use for standalone CLIP training, one-off evaluation or embedding, generic k-NN mining, or AOI/ChangeNet DEFT workflows.
NVIDIAOfficial
NVIDIA TAO Mask Auto-Label (MAL) tool for weakly-supervised segmentation using ViT-MAE backbones.
NVIDIAOfficial
Install and verify NVIDIA cuOpt for Python, C, or REST server environments.
NVIDIAOfficial
Orchestrates the creation and maintenance of an agent's ETHOS.md intent snapshot.
NVIDIAOfficial
Execute the end-to-end CDF AnomalyGen pipeline for synthetic defect image generation and model fine-tuning.
NVIDIAOfficial
Manage HOLOLINK_def.svh configuration files for HSB IP through generation, validation, and comparison workflows.
NVIDIAOfficial
NVIDIA DeepStream SDK development with Python pyservicemaker API. Use when building video analytics pipelines, GStreamer-based video processing, TensorRT inference integration, object detection/tracking, or Kafka/message broker integration.