agent-gpu-skills
GPU kernel development with CUDA, Triton, CUTLASS, and TileLang
All Skills in This Repository (4)
Pure Emerald Level Indicators
slowlyCtriton-skill
Write, debug, and optimize GPU kernels using Triton and Gluon DSLs.
slowlyCcuda-skill
Search local NVIDIA CUDA documentation for PTX and API references.
slowlyCcutlass-skill
Develop, debug, and optimize CUTLASS and CuTeDSL GPU kernels for NVIDIA architectures.
slowlyCsglang-skill
Navigate SGLang source code and documentation for development and optimization.
Frequently Asked Questions
FAQPage SchemaHow to install agent-gpu-skills?▼
Run `npx skills add slowlyC/agent-gpu-skills --all -g -y` in your terminal to install all four GPU skills globally.
What does agent-gpu-skills do?▼
It gives your coding assistant local access to official CUDA and PTX documentation plus searchable Triton, CUTLASS, and TileLang source code, so it can write and debug GPU kernels accurately.
Which GPU frameworks are covered?▼
The suite covers raw CUDA and PTX, Triton and Gluon, CUTLASS with CuTe and CuTeDSL, and TileLang, including NVIDIA profiling tools like Nsight Compute and Nsight Systems.
Does it work with Claude Code, Cursor, and Codex?▼
Yes. All skills follow the standard SKILL.md format and install into the skill directories of Cursor, Claude Code, Codex, Gemini CLI, and Qoder.
Do I need internet access to look up CUDA docs?▼
No. The CUDA skill ships local snapshots of the official PTX ISA, Runtime API, Driver API, and Nsight documentation, so lookups work offline.
Related Repositories in Software Engineering
View All in Software Engineering→
openclawopenclaw
Run a personal AI assistant across your devices and chat apps
obrasuperpowers
Gives coding agents a disciplined workflow from idea to merged code
facebookreact
AI agent skills for building, testing, and porting React core