Development · CUDA · For developers

Debug AOTInductor (AOTI) errors and crashes. An agent skill from pytorch/pytorch.

pytorch/pytorch

Debug hanging issues in SGLang distributed inference (TP/PP/DP/EP).

sgl-project/sglang

Create a CUDA Python pull request from an approved personal or organization-owned fork, including the GitHub CLI GraphQL fallback for renamed organization-owned forks.

NVIDIA/cuda-python

Explains why editing CUTLASS fused-MHA headers in ONNX Runtime can leave stale CUDA kernels after an incremental build, and how to force and verify a real rebuild.

microsoft/onnxruntime

把 LeetCUDA kernels/interview 教学源码（base/sgemv/sgemm/hgemm/flashattn/ffpaattn/fp8gemm.cuh/fp4gemm.cuh）+ ffpa-attn CuTe sm120 源码（csrc/cuffpa/cute fp8/fp4，commit 861d75e）写成中文 CUDA 技术书（6 Part 38 章 + CuTe…

xlite-dev/LeetCUDA

Builds ONNX Runtime from source with its build scripts, explaining the update, build and test phases, key flags and where the build output lands.

microsoft/onnxruntime

Condensed debugging method and tool recipes for Unix, Python and PyTorch programs: crashes, hangs, segfaults, wrong output, CUDA OOM, NaN values and slowness.

stas00/the-art-of-debugging

Builds ExecuTorch from source: the Python package, C++ runtime, model runners, Android and iOS cross-compilation and backend-specific builds, with environment checks.

pytorch/executorch

Find and safely apply per-operator refactoring / redundancy-reduction opportunities in a CV-CUDA operator (near-duplicate Tensor/VarShape kernels, reinvented shared utilities, dead code).

CVCUDA/CV-CUDA

Drafts ONNX Runtime release notes from commit history and contributor metadata using named presets for the full runtime or a scoped component.

microsoft/onnxruntime

A skill your agent uses for ANY bug, error, crash, wrong output, loss divergence, gradient explosion, test failure, CUDA error, distributed training hang, checkpoint load failure, or unexpected…

ByteDance-Seed/VeOmni

CUDA kernel development, debugging, and performance optimization for Claude Code.

technillogue/ptx-isa-markdown

Comprehensive C/C++ programming reference covering everything from C11-C23 and C++11-C++23, system programming, CUDA GPU computing, debugging tools, Rust interop, and advanced topics.

crazyguitar/cppcheatsheet

Analyzes Torch Profiler traces from SGLang, vLLM and TensorRT-LLM servers into kernel attribution, overlap and fusion tables.

BBuf/AI-Infra-Auto-Driven-SKILLS

A skill your agent uses when writing, debugging, porting, reviewing, or optimizing CUDA C++ or PTX kernels; investigating CUDA Runtime or Driver API behavior; profiling kernels with Nsight Systems…

vipshop/cache-dit

Generate a local phyai environment report for debugging system, Python, CUDA/GPU, dependency, workspace package, git, and PHYAI configuration issues.

mingti-org/phyai

Readable C/C++/Rust/CUDA code rules inspired by The Art of Readable Code.

crazyguitar/cppcheatsheet

Write, debug, and optimize TileLang kernels from local upstream language, JIT, autotuning, profiling, compiler, test, and example source.

slowlyC/agent-gpu-skills

A skill your agent uses when writing, debugging, porting, reviewing, or optimizing CUTLASS or CuTe C++ kernels and templates; navigating CUTLASS examples, collectives, epilogues, pipelines, GEMM…

vipshop/cache-dit

A skill your agent uses for performance profiling and optimization.

ByteDance-Seed/VeOmni

Adds verified layer guides such as L0 and L1 and compact GPU lanes to an existing Torch Profiler Chrome trace, changing how it looks but not how it ran.

BBuf/AI-Infra-Auto-Driven-SKILLS

Call this skill when you need to debug CUDA crashes in SGLang using kernel API logging

sgl-project/sglang

Draft, debug, and measure CUDA kernels and host launch workflows.

sablin39/tilelang-cuda-skills

Patterns and pitfalls for the ONNX-domain Attention operator's CUDA implementation in ONNX Runtime: dispatch cascade, eligibility limits, mask and bias kernels, and test routing.

microsoft/onnxruntime

Touch-lists for common OCANNL extension tasks: adding a primitive operation, adding or extending a backend, extending shape inference, and diagnosing output differences between backends.

ahrefs/ocannl

Review llama.cpp changes against project conventions and common reviewer pitfalls before a PR.

JakeATX/llamAmpere

Review 3DGS implementation code for correctness, performance bugs, and best practices.

jaccen/Awesome-Gaussian-Skills

A skill your agent uses when updating dependencies managed by uv: bumping a package version, upgrading the uv tool itself, updating torch/CUDA stack, switching transformers version, or regenerating…

ByteDance-Seed/VeOmni

Query current NVIDIA CUDA, PTX ISA, Runtime API, Driver API, Programming Guide, Best Practices, Nsight Compute, and Nsight Systems references.

slowlyC/agent-gpu-skills

Debug crashes and test failures via stack-traces, host/device logging, and DSL buffer inspection.

LuisaGroup/LuisaCompute

使用 PaddlePaddle 仓库规则评审 Pull Request 和全仓库代码变更，覆盖正确性、兼容性、算子、分布式、数值、性能、安全、测试、构建和 PR 信息。当需要审查 Paddle 的代码、测试、算子 YAML、C++/CUDA/XPU kernel、Python API、分布式逻辑或 CI 配置时使用。

PaddlePaddle/Paddle

Rigor Debug / Rigor Audit skill for deep learning research work.

lllllllama/RigorPilot-Skills

Measure TileLang kernel and workflow performance, choose benchmark or timeline tools, and compare implementations fairly.

sablin39/tilelang-cuda-skills

Reference guide for the MPK compilation-to-runtime pipeline.

mirage-project/mirage

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

matlab/matlab-agentic-toolkit

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

spiriMirror/libuipc

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

A skill your agent uses when adding, modifying, optimizing, or debugging CuTile autotuning code.

NVIDIA/skills

Diagnose CUDA "illegal instruction" / kernel crashes on Triton kernels that reference to TMA loads or stores (maketensordescriptor, TensorDescriptor, descriptor.load, descriptor.store…

facebookexperimental/triton

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

A skill your agent uses when the operator is authoring, building, loading, or debugging a custom doca-bench plug-in — a versioned shared library with DOCAEXPERIMENTAL-marked C entry points that…

NVIDIA/skills

A skill your agent uses when the user is measuring GPU-kernel-initiated RDMA WRITE latency through doca-gpunetio — building and running the gpunetioibwritelat client + server pair under…

NVIDIA/skills

How to create a pull request for the intel/torch-xpu-ops repository.

intel/torch-xpu-ops

How HOT-Step's custom flash-attention training ops (GGMLOPFLASHATTNTRAIN/BACK) work, what the AS1.5 DiT trainer campaign proved and disproved, and the exact contract for porting flash mode to the…

scragnog/HOT-Step-CPP

Evidence-gated workflow for MoE performance optimization in Megatron Bridge.

NVIDIA/skills

AI for Science 场景下的昇腾 NPU Profiling 采集与性能分析 Skill，用于在华为 Ascend NPU 上使用 torchnpu.profiler 采集 L0、L1、L2 级性能数据，分析训练或推理中的算子耗时、调用栈、内存与瓶颈，并指导后续调优。

ascend-ai-coding/awesome-ascend-skills

Install, manage, and run ComfyUI instances. An agent skill from sundial-org/awesome-openclaw-skills.

sundial-org/awesome-openclaw-skills

CUDA debugging skill for GPU program correctness. An agent skill from mohitmishra786/low-level-dev-skills.

mohitmishra786/low-level-dev-skills