NeVibeCode · repository · 312 items
skills: a set by nvidia
The nvidia/skills repository on GitHub: 312 cards with a description and an install command. Inside: 311 skills and 1 plugin. They install into Claude Code, Codex and Cursor.
About the repository
- GitHub
- nvidia/skills
- Stars
- ★ 4,106
- Installs across cards
- 334,469
- Licenses
- Apache-2.0, CC-BY-4.0, LicenseRef-Proprietary
- Source updated
- 2026-11-20
- In the catalog since
- 2026-10-25
nv-generate-vae-finetune skill nv-generate-vae-finetune is a Claude Code skill for preparing MAISI VAE fine-tuning with CT or MRI NIfTI volumes. It checks the datalist and file paths before training.
CX Codex +1
1,800installs
nv-segment-ct-finetune skill nv-segment-ct-finetune is a Claude Code skill that fine-tunes VISTA3D on CT images and labels in NIfTI format. It runs standard training or fixed-class softmax training.
CX Codex +1
1,800installs
nv-reason-cxr skill Runs NV-Reason-CXR-3B smoke tests on chest X-ray images. Useful for verifying pipeline command shapes and parsing JSON reasoning outputs.
CX Codex +1
1,800installs
nv-generate-ct-rflow skill nv-generate-ct-rflow is a Claude Code skill that orchestrates synthetic CT volume and mask generation via NV-Generate-CTMR rflow-ct. Executes inference wrappers and validates output packs.
CX Codex +1
1,800installs
nv-segment-ctmr skill Runs NV-Segment-CTMR on CT or MRI NIfTI volumes and writes a label map plus a JSON report. Not meant for clinical interpretation.
CX Codex +1
1,800installs
nv-generate-mr skill nv-generate-mr generates synthetic body MRI volumes using NV-Generate-CTMR rflow-mr. Useful for staging T1 MRI volume synthesis without paired segmentations.
CX Codex +1
1,800installs
nv-segment-ct skill nv-segment-ct — a skill that runs NVIDIA's VISTA3D model on CT NIfTI volumes and writes a label map plus a JSON evidence file. Supported across Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode.
CX Codex +1
1,800installs
nv-generate-mr-brain-finetune skill nv-generate-mr-brain-finetune finetunes the NV-Generate-CTMR MR-Brain v1 diffusion model on your T1, T2, FLAIR, SWI or MRA NIfTI volumes. A preflight pass runs first, without a GPU.
CX Codex +1
1,800installs
nv-generate-mr-brain skill nv-generate-mr-brain generates synthetic T1, T2, FLAIR, SWI, or MRA brain MRI volumes with NV-Generate-CTMR MR-Brain v1. It provides test NIfTI volumes without clinical scans.
CX Codex +1
1,800installs
tilegym-cutile-python skill tilegym-cutile-python turns the assistant into a cuTile specialist: writing, debugging and tuning GPU kernels in NVIDIA's Python DSL. Useful when porting tensor ops off PyTorch by hand.
CX Codex +1
1,700installs
tao-launch-workflow skill tao-launch-workflow — a skill for Claude Code, Codex, Cursor and more: a mandatory preflight gate before any TAO workflow fires. Triggers before training, AutoML, inference, export, and TensorRT engine generation.
CX Codex +1
1,700installs
tao-setup-nvidia-gpu-host skill tao-setup-nvidia-gpu-host is a skill that checks and installs the NVIDIA driver, CUDA Toolkit and NVIDIA Container Toolkit for Docker and Kubernetes. It prepares hosts for TAO GPU workflows.
CX Codex +1
1,700installs
tao-finetune-huggingface-model skill tao-finetune-huggingface-model is a skill that fine-tunes HuggingFace CV, VLM, and LLM models on local NVIDIA GPUs in an NGC PyTorch container when no dedicated TAO skill matches.
CX Codex +1
1,700installs
hsb-app skill The hsb-app skill finds Holoscan Sensor Bridge example apps on a connected devkit, filters them by your hardware, and runs one in the demo container.
CX Codex +1
1,700installs
hsb-flash skill hsb-flash flashes the FPGA on an HSB board attached to an NVIDIA devkit. Detects the board type first, then walks through each step with confirmation.
CX Codex +1
1,700installs
hsb-test skill hsb-test runs automated QA test plans against Holoscan Sensor Bridge hardware on a remote devkit. It filters tests from your document, executes them in a demo container, and writes a results report.
CX Codex +1
1,700installs
hsb-setup skill hsb-setup — a skill for Claude Code and five other platforms that takes the NVIDIA Holoscan Sensor Bridge from a bare devkit to a running demo container.
CX Codex +1
1,700installs
tilegym-cutile-autotuning skill tilegym-cutile-autotuning walks your agent through wiring autotuning into CuTile kernels: search once, cache the winner, launch straight from the cache.
CX Codex +1
1,600installs
tao-run-on-slurm skill tao-run-on-slurm — a skill for Claude Code, Codex, Cursor and other agents that submits TAO training, eval and inference jobs to a remote SLURM GPU cluster over SSH. Each job runs inside a Pyxis/Enroot container against Lustre-staged data.
CX Codex +1
1,600installs
tao-run-on-kubernetes skill tao-run-on-kubernetes turns a TAO spec-bundle into a Kubernetes Job and submits it with plain kubectl. For training runs and inference on EKS, GKE, AKS or an on-prem GPU cluster.
CX Codex +1
1,600installs
tilegym-improve-cutile-kernel-perf skill tilegym-improve-cutile-kernel-perf is a Claude Code skill for systematically tuning cuTile kernels in TileGym, identifying bottlenecks, and measuring latency on NVIDIA GPUs.
CX Codex +1
1,600installs
tilegym-monkey-patch-kernels-to-transformers skill The tilegym-monkey-patch-kernels-to-transformers skill wires TileGym kernels into Hugging Face Transformers models by swapping components before model creation. Use it to integrate existing kernels or build new cuTile ones.
CX Codex +1
1,600installs
tao-run-inference-service skill tao-run-inference-service deploys network-specific NVIDIA TAO inference microservices in containers and halts them. Use it to run checkpoints and query endpoints without cloud storage.
CX Codex +1
1,600installs
tao-validate-dataset-format skill The tao-validate-dataset-format skill runs tao-daft validate on NVIDIA TAO DAFT data and helps interpret the results. Use it to check dataset structure, schemas, and cross-references.
CX Codex +1
1,600installs
tao-train-mask-auto-encoder skill tao-train-mask-auto-encoder — a skill for pretraining and fine-tuning Masked Auto-Encoders in NVIDIA TAO. Lets an agent drive train, evaluate, inference and export for MAE backbones.
CX Codex +1
1,600installs
tao-train-single-step skill tao-train-single-step runs standard train, eval, and export workflows for NVIDIA TAO models without AutoML loops. Built for single-run fine-tuning on labeled datasets.
CX Codex +1
1,600installs
tao-train-centerpose skill tao-train-centerpose is a Claude Code skill for training CenterPose models to detect keypoints and estimate 6-DoF object poses, with workflows for evaluation, inference, and export.
CX Codex +1
1,600installs
tao-run-automl-deft-pipeline skill tao-run-automl-deft-pipeline — a skill that chains three training phases: hyperparameter search, the DEFT loop, and a second AutoML pass on the augmented dataset.
CX Codex +1
1,600installs
tao-train-metric-learning-recognition skill tao-train-metric-learning-recognition is a Claude Code skill that trains, evaluates, and exports NVIDIA TAO metric learning models for visual retrieval and fine-grained matching.
CX Codex +1
1,600installs
tao-convert-dataset-format skill The tao-convert-dataset-format skill runs tao-daft convert to move NVIDIA TAO DAFT datasets between supported formats. Use it when preparing DAFT data for VLM training.
CX Codex +1
1,600installs
tao-generate-video-reasoning-annotations skill tao-generate-video-reasoning-annotations turns raw videos into a training dataset with captions, descriptions and QA pairs carrying reasoning traces. Useful when you need labelled video data.
CX Codex +1
1,600installs
tao-list-capabilities skill tao-list-capabilities — a skill for Claude Code and Cursor: builds a list of TAO Skill Bank capabilities from manifests — apps, models, execution platforms, AutoML. Answers queries about available models, workflows, and AutoML support.
CX Codex +1
1,600installs
tao-run-automl skill The tao-run-automl skill runs AutoML and hyperparameter search for NVIDIA TAO models. Use it to tune training or another supported model action on a selected platform.
CX Codex +1
1,600installs
tao-finetune-cosmos-reason skill The tao-finetune-cosmos-reason skill plans and runs Cosmos3 fine-tuning through Cosmos Framework or Cosmos-RL, validating every input before training starts.
CX Codex +1
1,600installs
tao-route-visual-changenet-samples skill tao-route-visual-changenet-samples splits weak VCN samples into Mining and AnomalyGen subsets after gap analysis. Use it for the routing step in an SDA iteration.
CX Codex +1
1,600installs
tao-train-fast-foundation-stereo skill tao-train-fast-foundation-stereo is a skill for training and evaluating FastFoundationStereo, exporting the model and running inference to produce disparity maps from stereo pairs.
CX Codex +1
1,600installs
tao-train-sparse4d skill tao-train-sparse4d prepares calibrated multi-camera data and trains, evaluates, or quantizes Sparse4D 3D perception models in NVIDIA TAO.
CX Codex +1
1,600installs
tao-train-foundation-stereo skill tao-train-foundation-stereo is a Claude Code skill for training and running FoundationStereo on stereo image pairs, producing disparity maps for 3D reconstruction.
CX Codex +1
1,600installs
tao-train-oneformer skill tao-train-oneformer is a Claude Code skill that trains, evaluates and exports OneFormer models in TAO for panoptic, instance and semantic segmentation.
CX Codex +1
1,600installs
tao-train-deformable-detr skill tao-train-deformable-detr is a skill for configuring 2D object detection workflows with Deformable DETR in NVIDIA TAO Toolkit. It supports training, evaluating, quantizing, and exporting models.
CX Codex +1
1,600installs
81–120 of 312