CUDA agent skills, page 5

Review a CV-CUDA operator's BENCHMARK coverage — drivers, layout axis, baselines, the basic-tier floor, row counts, and coverage statistics.

CVCUDA/CV-CUDA

Optimize MATLAB design files for GPU Coder to generate faster CUDA code.

matlab/matlab-agentic-toolkit

pyqula's CPU/GPU switch (src/pyqula/gpu.py), how a routine is routed onto the device, per-call precision, and the tiered porting plan in documentation/gpuportingplan.md.

joselado/pyqula

给 FlashRec 引擎接入一个新模型架构（新的 HF checkpoint / 非 Qwen3 结构）。涵盖模型定义、权重合并加载、FP8 双路径、融合 kernel 接线、CUDA graph 兼容、精度校验、以及压测+trace 验证闭环。当用户要"增加/支持/接入新模型"时使用。

sohu-mptc/FlashRec

Step-by-step tutorial for adding a new lightweight JIT CUDA kernel to sglang.kernels JIT infrastructure and public operator groups

sgl-project/sglang

Review a CV-CUDA operator's DOCS & API artifacts — operatorlist row, Python autofunction (fn + into), Limitations-table-vs-code consistency, docstrings, and SPDX headers.

CVCUDA/CV-CUDA

Builds the code for a frozen research experiment test-first, with leakage controls, seed handling and saved evidence so results can be rerun and audited.

Light0305/Light-skills

Opinionated app components building on top of ./ui primitives

JakeATX/llamAmpere

Start, validate, debug, and stop an AReno OpenAI-compatible serving endpoint.

inclusionAI/AReno

Deploy vLLM using Docker (pre-built images or build-from-source) with NVIDIA GPU support and run the OpenAI-compatible server.

vllm-project/vllm-skills

GPU-accelerates scientific Python on NVIDIA hardware and verifies that the result is correct and faster.

K-Dense-AI/scientific-agent-skills

Builds and troubleshoots TorchDrug 0.2.1 workflows for molecular graphs, property prediction, self-supervised pretraining, molecule generation, retrosynthesis, protein representation learning, and…

K-Dense-AI/scientific-agent-skills

Review a CV-CUDA operator's input-type, layout, dtype, and channel support matrix.

CVCUDA/CV-CUDA

Overview of the main directories and important files in the repository.

spiriMirror/libuipc

Write code that calls the installed nvMolKit Python API for GPU-accelerated, batched RDKit-style operations - Morgan fingerprints, Tanimoto/cosine similarity, ETKDG conformer embedding, MMFF/UFF…

NVIDIA-BioNeMo/bionemo-agent-toolkit

Detect out-of-bounds memory accesses in CPU or GPU code using static interval analysis and runtime assertions/printfs.

ROCm/FlyDSL

A skill your agent uses when running or debugging interactive Skippy prompts against staged serving, including lab sync, native builds, stage startup, the HTTP prompt REPL, and process lifecycle.

Mesh-LLM/mesh-llm

Update tools/scripts/generatebinarybuildmatrix.py when a PyTorch release goes live.

pytorch/test-infra

Review a CV-CUDA operator's test coverage, including C++ correctness, required cross-layout parity, correctness rigor, and the Python API surface.

CVCUDA/CV-CUDA

Apply Gkeyll naming conventions when creating, editing, or reviewing C, CUDA, or Lua files and code elements.

gkeyllorg/gkeyll

Deploy and manage vLLM for high-throughput LLM inference. An agent skill from sickn33/agentic-awesome-skills.

sickn33/agentic-awesome-skills

A skill your agent uses when the user needs to choose the tile, dtype, ctagroup::1/2, SMEM operand layout, or TMEM accumulator mapping for a tcgen05 MMA on B200/Blackwell, or to implement an…

mirage-project/mirage

A skill your agent uses when a B200/TIRx/CUDA warp-specialized kernel fails to compile, deadlocks, hits an illegal memory access, produces wrong results, or is "correct but slow".

mirage-project/mirage

A skill your agent uses when the user wants to confirm whether an existing CUDA extension/binary can run on B200, configure compute100/sm100 or the architecture-specific sm100a, or check PTX/cubin…

mirage-project/mirage

A skill your agent uses when you need a NEW or optimized MPK CUDA kernel (a per-task .cuh under include/mirage/persistentkernel/tasks/) that must PROVABLY beat a target WITHOUT the kernel-optimizer…

mirage-project/mirage

Use FP16/BF16 mixed precision to accelerate training and reduce memory.

aiming-lab/AutoResearchClaw

AI demos and GPU compute with Gradio Spaces and Hugging Face Spaces ZeroGPU.

sickn33/agentic-awesome-skills

Install Triton + SageAttention to accelerate ComfyUI (the sageattn attentionmode and inductor torch.compile used by WanVideoWrapper / many video graphs).

artokun/comfyui-mcp

Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13).

wshobson/agents

A skill your agent uses when creating, migrating, or debugging pixi environments, especially for scientific Python, bioinformatics, single-cell analysis, CUDA/PyTorch, Jupyter/VS Code kernels…

xuzhougeng/wisp-science

Write, review, and run high-level xTBloom Python GFN2-xTB inference with Calculator, Structure, and BatchCalculator, including single systems, repeated geometry updates, heterogeneous ragged…

jinzhezenggroup/computational-chemistry-agent-skills

Install or verify the correct ONNX Runtime build (and the onnx package) for a user's accelerator backend before Quark's ONNX-to-ONNX flow.

amd/Quark

Check whether a CV-CUDA operator is READY to optimize (correctness + bench coverage + captured baseline + profiling) per .agents/guidance/OPTIMIZATIONGUIDELINES.md.

CVCUDA/CV-CUDA

5-stage kernel correctness verification protocol for Triton and CUDA kernels.

ZJLi2013/awesome-kernel-skills

XMake build configuration, options, commands, and patterns for LuisaCompute.

LuisaGroup/LuisaCompute

HeartMuLa: Suno-like song generation from lyrics + tags. An agent skill from Luciole-Studio/Misaka-Agent.

Luciole-Studio/Misaka-Agent

Continue pretraining from an existing KERMT checkpoint. An agent skill from NVIDIA-BioNeMo/bionemo-agent-toolkit.

NVIDIA-BioNeMo/bionemo-agent-toolkit

Set up and run HeartMuLa, the open-source music generation model family (Suno-like).

RedWoodOG/Hermes-Desktop

Inspect a target ONNX model and prepare metadata for Quark ONNX PTQ planning.

amd/Quark

Mandatory pre-flight compute resource check before running experiments.

OpenLAIR/dr-claw

Maps every HOT-Step CPP feature to its route file, service, UI folder, and engine subsystem, including port topology and the browser-to-engine request path.

scragnog/HOT-Step-CPP

Build a Quark ONNX PTQ quantization plan from modelanalysis.json and user intent.

amd/Quark

Diagnose failed Quark installation, PTQ execution, script generation, or export attempts.

amd/Quark

Install or verify the correct PyTorch build for a user's accelerator backend before Quark installation.

amd/Quark

Diagnoses and resolves MCP server registration failures, GPU detection, BigQuery authentication, index build failures, import errors, search quality issues, and performance problems.

RobThePCGuy/Claude-Patent-Creator

Generate C/C++ or CUDA code from an AI model (PyTorch, LiteRT) using MATLAB Coder or GPU Coder.

matlab/matlab-agentic-toolkit

Generate, verify, refine, and accelerate C/C++ or CUDA code from MATLAB with MATLAB Coder, Embedded Coder, GPU Coder, or MATLAB Test.

matlab/matlab-agentic-toolkit

Build and run LAMMPS molecular dynamics with isolated MLIP-specific binaries (MACE, MatGL/CHGNet, FairChem) to avoid Python and Torch stack conflicts.

learningmatter-mit/AtomisticSkills