Updated 3mo ago
by tensormux
kernel-skills
Open source skill library for AI coding agents to write, optimize, and debug high performance compute kernels across CUDA, Triton, and quantized workloads.
82 GitHub stars (82)
+2 in 16 days
| Period | New stars |
|---|---|
| Oct 1 | 0 |
| Oct 2 | 0 |
| Oct 3 | 0 |
| Oct 4 | 0 |
| Oct 5 | 0 |
| Oct 6 | 0 |
| Oct 7 | −1 |
| Oct 8 | 0 |
| Oct 9 | 0 |
| Oct 10 | 0 |
| Oct 11 | 0 |
| Oct 12 | 0 |
| Oct 13 | 0 |
| Oct 14 | +1 |
| Oct 15 | +1 |
| Oct 16 | +1 |
- 30 skills (30 skills found in its SKILL.md files)
- 10 forks (10 forks)
- 3mo ago last update (Last updated 3mo ago)
Install
Skills CLI
npx skills add tensormux/kernel-skills
Asks which skills and agents to install. Just one: npx skills add tensormux/kernel-skills --skill avoid-warp-divergence — add -g for every project.
Claude Code
git clone https://github.com/tensormux/kernel-skillscp -r "kernel-skills/skills/cuda/avoid-warp-divergence" ~/.claude/skills/
Copies the avoid-warp-divergence skill into ~/.claude/skills/ for every project (use .claude/skills/ for one project). Repeat for the other skills listed below.
Plugin
/plugin marketplace add tensormux/kernel-skills
Run inside Claude Code if the repository ships a plugin marketplace, then pick plugins from /plugin.
Skills can run code on your machine — read the SKILL.md first.
What it does
In the author's own words, from each skill's SKILL.md — showing the first 30 of 35 files.
-
avoid-warp-divergence
No description in the SKILL.md.
skills/cuda/avoid-warp-divergence/SKILL.md on GitHub (opens in a new tab) -
choose-launch-configuration
No description in the SKILL.md.
skills/cuda/choose-launch-configuration/SKILL.md on GitHub (opens in a new tab) -
debug-cuda-kernel-correctness
No description in the SKILL.md.
skills/cuda/debug-cuda-kernel-correctness/SKILL.md on GitHub (opens in a new tab) -
optimize-global-memory-access
No description in the SKILL.md.
skills/cuda/optimize-global-memory-access/SKILL.md on GitHub (opens in a new tab) -
optimize-shared-memory-tiling
No description in the SKILL.md.
skills/cuda/optimize-shared-memory-tiling/SKILL.md on GitHub (opens in a new tab) -
write-cuda-gemm-kernel
No description in the SKILL.md.
skills/cuda/write-cuda-gemm-kernel/SKILL.md on GitHub (opens in a new tab) -
write-cuda-layernorm-kernel
No description in the SKILL.md.
skills/cuda/write-cuda-layernorm-kernel/SKILL.md on GitHub (opens in a new tab) -
write-cuda-reduction-kernel
No description in the SKILL.md.
skills/cuda/write-cuda-reduction-kernel/SKILL.md on GitHub (opens in a new tab) -
write-cuda-softmax-kernel
No description in the SKILL.md.
skills/cuda/write-cuda-softmax-kernel/SKILL.md on GitHub (opens in a new tab) -
optimize-prefill-vs-decode-kernels
No description in the SKILL.md.
skills/inference/optimize-prefill-vs-decode-kernels/SKILL.md on GitHub (opens in a new tab) -
write-tensorrt-plugin-integration-plan
No description in the SKILL.md.
skills/inference/write-tensorrt-plugin-integration-plan/SKILL.md on GitHub (opens in a new tab) -
write-triton-dequant-kernel
No description in the SKILL.md.
skills/inference/write-triton-dequant-kernel/SKILL.md on GitHub (opens in a new tab) -
write-triton-fused-add-rmsnorm-kernel
No description in the SKILL.md.
skills/inference/write-triton-fused-add-rmsnorm-kernel/SKILL.md on GitHub (opens in a new tab) -
write-triton-kv-cache-append-kernel
No description in the SKILL.md.
skills/inference/write-triton-kv-cache-append-kernel/SKILL.md on GitHub (opens in a new tab) -
write-triton-rmsnorm-kernel
No description in the SKILL.md.
skills/inference/write-triton-rmsnorm-kernel/SKILL.md on GitHub (opens in a new tab) -
write-triton-rope-kernel
No description in the SKILL.md.
skills/inference/write-triton-rope-kernel/SKILL.md on GitHub (opens in a new tab) -
write-triton-sampling-kernel
No description in the SKILL.md.
skills/inference/write-triton-sampling-kernel/SKILL.md on GitHub (opens in a new tab) -
write-triton-silu-mul-kernel
No description in the SKILL.md.
skills/inference/write-triton-silu-mul-kernel/SKILL.md on GitHub (opens in a new tab) -
write-vllm-custom-op-integration-plan
No description in the SKILL.md.
skills/inference/write-vllm-custom-op-integration-plan/SKILL.md on GitHub (opens in a new tab) -
choose-tile-size-and-work-partitioning
No description in the SKILL.md.
skills/patterns/choose-tile-size-and-work-partitioning/SKILL.md on GitHub (opens in a new tab) -
fuse-elementwise-ops
No description in the SKILL.md.
skills/patterns/fuse-elementwise-ops/SKILL.md on GitHub (opens in a new tab) -
handle-boundary-conditions
No description in the SKILL.md.
skills/patterns/handle-boundary-conditions/SKILL.md on GitHub (opens in a new tab) -
write-kernel-test-plan
No description in the SKILL.md.
skills/patterns/write-kernel-test-plan/SKILL.md on GitHub (opens in a new tab) -
write-numerically-stable-kernel
No description in the SKILL.md.
skills/patterns/write-numerically-stable-kernel/SKILL.md on GitHub (opens in a new tab) -
port-cuda-kernel-to-hip
No description in the SKILL.md.
skills/portability/port-cuda-kernel-to-hip/SKILL.md on GitHub (opens in a new tab) -
port-cuda-kernel-to-triton
No description in the SKILL.md.
skills/portability/port-cuda-kernel-to-triton/SKILL.md on GitHub (opens in a new tab) -
write-backend-agnostic-kernel-plan
No description in the SKILL.md.
skills/portability/write-backend-agnostic-kernel-plan/SKILL.md on GitHub (opens in a new tab) -
debug-quantized-kernel-accuracy
No description in the SKILL.md.
skills/quantization/debug-quantized-kernel-accuracy/SKILL.md on GitHub (opens in a new tab) -
write-fp8-kernel
No description in the SKILL.md.
skills/quantization/write-fp8-kernel/SKILL.md on GitHub (opens in a new tab) -
write-int8-quantized-kernel
No description in the SKILL.md.
skills/quantization/write-int8-quantized-kernel/SKILL.md on GitHub (opens in a new tab)
No skill matches that filter.
Review before you install
A skill is a set of instructions and scripts your agent will follow, and it can run code on your machine. Read the SKILL.md and any scripts first — this listing is not a security review.
Similar skills
-
agentic-awesome-skillsby sickn33 AAS Core is the local, agent-first control plane for complete catalog discovery, agent-owned selection, stack validation, and planning, bac…
-
vibe-coding-cnby tradecatlabs Vibe Coding 从入门到精通教程｜AI 结对编程工作流｜Prompt、Skill、Workflow、上下文管理、codex实战指南
-
autoharnessby tigerless-labs Autoharness — a self-learning skill layer for Claude Code — distills skills from your real sessions, updates them as you work, and prunes t…
-
ccpmby automazeio Project management skill system for Agents that uses GitHub Issues and Git worktrees for parallel agent execution.
-
superpowers-zhby jnMetaCode 🦸 AI 编程超能力 · 中文增强版 — superpowers（250k+ ⭐）完整汉化 + 4 个中国原创 skills，让 Claude Code / Copilot CLI / Hermes Agent / Cursor / Windsurf / Kiro / Gemi…
-
gentle-aiby Gentleman-Programming Gentle-AI configures the AI coding agents you already use: Claude Code, Cursor, OpenCode, Codex, Pi, and more. Choose persistent memory, Or…