Debug Distributed Hang
sgl-project/sglang
Debug hanging issues in SGLang distributed inference (TP/PP/DP/EP).
CUDA kernel development, debugging, and performance optimization for Claude Code.
Install Cuda
$ npx skills add technillogue/ptx-isa-markdown --skill cuda -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install technillogue/ptx-isa-markdown cuda --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/technillogue/ptx-isa-markdown.git skills-src && mkdir -p .claude/skills && cp -r skills-src/cuda_skill .claude/skills/cuda && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "cuda" agent skill from https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skill into .claude/skills/cuda/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cuda", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skillType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add technillogue/ptx-isa-markdown --skill cuda -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install technillogue/ptx-isa-markdown cuda --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/technillogue/ptx-isa-markdown.git skills-src && mkdir -p .agents/skills && cp -r skills-src/cuda_skill .agents/skills/cuda && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "cuda" agent skill from https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skill into .agents/skills/cuda/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cuda", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add technillogue/ptx-isa-markdown --skill cuda -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install technillogue/ptx-isa-markdown cuda --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/technillogue/ptx-isa-markdown.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/cuda_skill .cursor/skills/cuda && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "cuda" agent skill from https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skill into .cursor/skills/cuda/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cuda", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/technillogue/ptx-isa-markdown.git --path cuda_skill--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add technillogue/ptx-isa-markdown --skill cuda -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install technillogue/ptx-isa-markdown cuda --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/technillogue/ptx-isa-markdown.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/cuda_skill .gemini/skills/cuda && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "cuda" agent skill from https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skill into .gemini/skills/cuda/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cuda", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install technillogue/ptx-isa-markdown cudaInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add technillogue/ptx-isa-markdown --skill cuda -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/technillogue/ptx-isa-markdown.git skills-src && mkdir -p .github/skills && cp -r skills-src/cuda_skill .github/skills/cuda && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "cuda" agent skill from https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skill into .github/skills/cuda/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cuda", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add technillogue/ptx-isa-markdown --skill cuda -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install technillogue/ptx-isa-markdown cuda --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/technillogue/ptx-isa-markdown.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/cuda_skill .opencode/skills/cuda && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "cuda" agent skill from https://github.com/technillogue/ptx-isa-markdown/tree/main/cuda_skill into .opencode/skills/cuda/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "cuda", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
cuda- GitHub stars
- 229
- Token cost
- ~2.5k tokens
- SKILL.md length
- 828 words
- Files
- 665 (incl. references)
- Skills in repo
- 1
- Repo updated
- First seen
- Licence
- None found
At a glance
CUDA kernel development, debugging, and performance optimization for Claude Code.
- Works in 5 steps: Reproduce minimally — Isolate the… → Add printf — Before any tool, add printf… → Run compute-sanitizer — Catch memory… → …
- Optimizing CUDA code
- SKILL.md covers Core Philosophy, Debugging Workflow, Performance Optimization… and Compilation Reference, plus 3 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Cuda is an agent skill from technillogue/ptx-isa-markdown. CUDA kernel development, debugging, and performance optimization for Claude Code. Use when writing, debugging, or optimizing CUDA code, GPU kernels, or parallel algorithms. Covers non-interactive profiling with nsys/ncu, debugging with cuda-gdb/compute-sanitizer, binary inspection with cuobjdump, and performance analysis workflows. Triggers on CUDA, GPU programming, kernel optimization, nsys, ncu, cuda-gdb, compute-sanitizer, PTX, GPU profiling, parallel performance.
Its SKILL.md is about 2.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 667 other files, including reference files (for example `references/cuda-driver-docs/INDEX.md`, `references/cuda-driver-docs/data-structures/structcu__dev__sm__resource__group__params.md` and `references/cuda-driver-docs/data-structures/structcuaccesspolicywindow__v1.md`).
It sits in Development, covering Performance optimization, GPU and accelerator computing and Debugging. It works with CUDA. The repository describes itself as: PTX ISA 9.1 documentation converted to searchable markdown. Includes Claude Code skill for CUDA development.
When your agent uses it
- Optimizing CUDA code
- Parallel algorithms
- GPU programming
- Kernel optimization
Example prompts
- “/cuda”
Workflow steps
5 steps, taken from the first numbered list in SKILL.md.
- Reproduce minimally — Isolate the failing kernel with smallest possible input
- Add printf — Before any tool, add printf in device code to trace execution
- Run compute-sanitizer — Catch memory errors non-interactively
- If still stuck, try cuda-gdb non-interactively for backtrace
- When tools fail — Minimize the diff between working and broken code. Read it. The bug is in the diff.
What it can do on your machine
Read from SKILL.md and the folder at commit 64cfba5. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md (its code samples are bash and cuda).
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Cuda loads about 2.5k tokens when it runs, and up to ~809k if it reads all its reference files. Until then it costs about 119 tokens; SKILL.md has 828 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
Without a licence we can't republish the file, so here is its outline and opening line. It has 828 words (~2,491 tokens).
“Measure before guessing. GPU performance is deeply counterintuitive. Profile first, hypothesize second, change third, verify fourth.”
- name
- cuda
Files
SKILL.md and 664 other files (references) in cuda_skill of technillogue/ptx-isa-markdown.
- SKILL.md
- references/cuda-driver-docs/INDEX.md
- references/cuda-driver-docs/data-structures/structcu__dev__sm__resource__group__params.md
- references/cuda-driver-docs/data-structures/structcuaccesspolicywindow__v1.md
- references/cuda-driver-docs/data-structures/structcuarraymapinfo__v1.md
- references/cuda-driver-docs/data-structures/structcuasyncnotificationinfo.md
- references/cuda-driver-docs/data-structures/structcucheckpointcheckpointargs.md
- references/cuda-driver-docs/data-structures/structcucheckpointgpupair.md
- references/cuda-driver-docs/data-structures/structcucheckpointlockargs.md
- references/cuda-driver-docs/data-structures/structcucheckpointrestoreargs.md
- references/cuda-driver-docs/data-structures/structcucheckpointunlockargs.md
- references/cuda-driver-docs/data-structures/structcuctxcigparam.md
- references/cuda-driver-docs/data-structures/structcuctxcreateparams.md
- references/cuda-driver-docs/data-structures/structcuda__array3d__descriptor__v2.md
- references/cuda-driver-docs/data-structures/structcuda__array__descriptor__v2.md
- references/cuda-driver-docs/data-structures/structcuda__array__memory__requirements__v1.md
- references/cuda-driver-docs/data-structures/structcuda__array__sparse__properties__v1.md
- references/cuda-driver-docs/data-structures/structcuda__batch__mem__op__node__params__v1.md
- … and 647 more
Open the folder on GitHubat commit 64cfba5
Compare with similar skills
Cuda next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Cuda this skilltechnillogue/ptx-isa-markdown | 229 | — | ~2.5k | Automated safety check: Pass | None | |
| Debug Distributed Hangsgl-project/sglang | 37k | 2 repos | ~2.4k | Automated safety check: Pass | Apache-2.0 | |
| CUTLASS FMHA Incremental Rebuildmicrosoft/onnxruntime | 22k | — | ~1.3k | Automated safety check: Pass | MIT | |
| The Art of Debuggingstas00/the-art-of-debugging | 1.7k | — | ~6.1k | Automated safety check: Notes | CC-BY-SA-4.0 | |
| Torch Profiler Layer TrackBBuf/AI-Infra-Auto-Driven-SKILLS | 925 | — | ~2k | Automated safety check: Pass | None | |
| Cppcrazyguitar/cppcheatsheet | 290 | — | ~1.8k | Automated safety check: Pass | MIT |
Similar skills
- Official
CUTLASS FMHA Incremental Rebuild
microsoft/onnxruntime
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.
The Art of Debugging
stas00/the-art-of-debugging
Condensed debugging method and tool recipes for Unix, Python and PyTorch programs: crashes, hangs, segfaults, wrong output, CUDA OOM, NaN values and slowness.
Torch Profiler Layer Track
BBuf/AI-Infra-Auto-Driven-SKILLS
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.
Cpp
crazyguitar/cppcheatsheet
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.
LLM Torch Profiler Trace Analysis
BBuf/AI-Infra-Auto-Driven-SKILLS
Analyzes Torch Profiler traces from SGLang, vLLM and TensorRT-LLM servers into kernel attribution, overlap and fusion tables.
Related
Works with
Categories
Questions about Cuda
What does Cuda do?
CUDA kernel development, debugging, and performance optimization for Claude Code. Cuda is an agent skill from technillogue/ptx-isa-markdown. CUDA kernel development, debugging, and performance optimization for Claude Code.
When should I use Cuda?
Cuda fits situations like: optimizing CUDA code; parallel algorithms; GPU programming; kernel optimization.
How do I install Cuda in Claude Code?
Run `npx skills add technillogue/ptx-isa-markdown --skill cuda -a claude-code`. Or copy the skill folder (cuda_skill in technillogue/ptx-isa-markdown) into .claude/skills/cuda in your project. Claude Code loads it when a task matches its description.
How do I install Cuda in Codex?
Run `npx skills add technillogue/ptx-isa-markdown --skill cuda -a codex`. Or copy the skill folder (cuda_skill in technillogue/ptx-isa-markdown) into .agents/skills/cuda in your project. Codex loads it when a task matches its description.
Can I use Cuda in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add technillogue/ptx-isa-markdown --skill cuda -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/cuda, .gemini/skills/cuda, .github/skills/cuda and .opencode/skills/cuda in your project.
What does Cuda need to run?
SKILL.md names no scripts, command-line tools or credentials: Cuda is instructions for the agent only.
Does Cuda access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Cuda safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Cuda use?
No licence was found for Cuda or its repository. Without one, default copyright applies: ask the author before reusing or redistributing it.
How many tokens does Cuda use?
About 2.5k tokens (SKILL.md is roughly 10k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 806k tokens, read only when the agent opens those files.
What are the alternatives to Cuda?
Skills that share tags, products or a category with Cuda: Debug Distributed Hang (sgl-project/sglang, 37k stars), CUTLASS FMHA Incremental Rebuild (microsoft/onnxruntime, 22k stars), The Art of Debugging (stas00/the-art-of-debugging, 1.7k stars) and Torch Profiler Layer Track (BBuf/AI-Infra-Auto-Driven-SKILLS, 925 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Cuda?
technillogue (a GitHub user) maintains it in technillogue/ptx-isa-markdown, which has 229 GitHub stars. The repository was last updated on December 24, 2025.
Source: technillogue/ptx-isa-markdown on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.