Aoti Debug
pytorch/pytorch
Debug AOTInductor (AOTI) errors and crashes. An agent skill from pytorch/pytorch.
A skill your agent uses whenever the user asks to debug something, fix a bug, troubleshoot an issue, or mentions "scientific debugging".
Install Scientific Debug
$ npx skills add UKGovernmentBEIS/vllm-lens --skill scientific-debug -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install UKGovernmentBEIS/vllm-lens scientific-debug --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/UKGovernmentBEIS/vllm-lens.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/scientific-debug .claude/skills/scientific-debug && 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 "scientific-debug" agent skill from https://github.com/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debug into .claude/skills/scientific-debug/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "scientific-debug", 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/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debugType 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 UKGovernmentBEIS/vllm-lens --skill scientific-debug -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install UKGovernmentBEIS/vllm-lens scientific-debug --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/vllm-lens.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/scientific-debug .agents/skills/scientific-debug && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "scientific-debug" agent skill from https://github.com/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debug into .agents/skills/scientific-debug/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "scientific-debug", 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 UKGovernmentBEIS/vllm-lens --skill scientific-debug -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install UKGovernmentBEIS/vllm-lens scientific-debug --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/vllm-lens.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/scientific-debug .cursor/skills/scientific-debug && 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 "scientific-debug" agent skill from https://github.com/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debug into .cursor/skills/scientific-debug/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "scientific-debug", 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/UKGovernmentBEIS/vllm-lens.git --path .claude/skills/scientific-debug--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 UKGovernmentBEIS/vllm-lens --skill scientific-debug -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install UKGovernmentBEIS/vllm-lens scientific-debug --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/vllm-lens.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/scientific-debug .gemini/skills/scientific-debug && 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 "scientific-debug" agent skill from https://github.com/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debug into .gemini/skills/scientific-debug/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "scientific-debug", 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 UKGovernmentBEIS/vllm-lens scientific-debugInstalls 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 UKGovernmentBEIS/vllm-lens --skill scientific-debug -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/vllm-lens.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/scientific-debug .github/skills/scientific-debug && 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 "scientific-debug" agent skill from https://github.com/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debug into .github/skills/scientific-debug/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "scientific-debug", 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 UKGovernmentBEIS/vllm-lens --skill scientific-debug -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install UKGovernmentBEIS/vllm-lens scientific-debug --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/vllm-lens.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/scientific-debug .opencode/skills/scientific-debug && 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 "scientific-debug" agent skill from https://github.com/UKGovernmentBEIS/vllm-lens/tree/main/.claude/skills/scientific-debug into .opencode/skills/scientific-debug/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "scientific-debug", 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
scientific-debug- GitHub stars
- 132
- Token cost
- ~556 tokens
- SKILL.md length
- 331 words
- Files
- 1
- Skills in repo
- 1
- Repo updated
- First seen
- Licence
- MIT
At a glance
A skill your agent uses whenever the user asks to debug something, fix a bug, troubleshoot an issue, or mentions "scientific debugging".
- Works in 4 steps: Read the evidence → Get more information → Reassess → …
- The user asks to debug something
- SKILL.md covers Step 1: Read the evidence, Step 2: Get more information, Step 3: Reassess and Step 4: Verify the fix
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Scientific Debug is an agent skill from UKGovernmentBEIS/vllm-lens. Use this skill whenever the user asks to debug something, fix a bug, troubleshoot an issue, or mentions "scientific debugging". Any debugging request should use this skill.
Its SKILL.md is about 560 tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Development, covering Debugging. The repository describes itself as: Extract residual-stream activations and apply steering vectors (including activation oracles) to any vLLM model during inference. The licence is MIT.
When your agent uses it
- The user asks to debug something
- Troubleshoot an issue
- Mentions scientific debugging
Example prompts
- “scientific debugging”
- “/scientific-debug”
Workflow steps
4 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 3d11fb0. 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.
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
Scientific Debug loads about 556 tokens when it runs. Until then it costs about 47 tokens; SKILL.md has 331 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
The full file from UKGovernmentBEIS/vllm-lens at commit 3d11fb0, republished under its MIT licence (© UKGovernmentBEIS). 331 words, ~556 tokens.
Download SKILL.mdSave it as
.claude/skills/scientific-debug/SKILL.md (or your agent's skills folder).- name
- scientific-debug
- description
- Use this skill whenever the user asks to debug something, fix a bug, troubleshoot an issue, or mentions "scientific debugging". Any debugging request should use this skill.
Scientific Debugging
Stop. Do not try random fixes. Your goal is to understand the root cause before changing anything.
Step 1: Read the evidence
Read the error, logs, and relevant code. If the cause is genuinely obvious from what you can see — state specifically what you believe is wrong, why you're confident, fix it, and you're done.
But be honest with yourself: RL training has made you overconfident. You often feel sure but are wrong, leading to a cycle of random patches that bloat the codebase. If you're not >95% certain of the root cause, do NOT attempt a fix. Go to Step 2.
Step 2: Get more information
You need more logging. This is the most important step — almost any bug becomes obvious with enough visibility.
- Increase log verbosity: set debug log levels, add env vars that enable verbose output, import loggers and set their level.
- If you're unsure how to get more logging for the specific library/framework involved, search the internet for how to enable debug logging for that tool. Don't guess from potentially stale knowledge.
- Add targeted print/log statements around the suspicious area.
- If it's not a Slurm/remote job, use breakpoints and a debugger.
Run the code again with enhanced logging.
Step 3: Reassess
With the new logs, are you now >95% confident in the root cause?
- Yes: State the cause specifically, explain why you're confident, and make the fix.
- No: Go back to Step 2. Dig deeper — read the source code of underlying libraries, find what env vars or config options control logging, check if you can set the logger's level programmatically. You can always get enough information to understand what's happening. Do not skip this and guess.
Repeat Steps 2-3 until you truly understand the problem. Only then fix it.
Step 4: Verify the fix
After applying your fix, rerun the code to confirm the bug is actually gone. If the job is expensive (large Slurm jobs, long GPU runs, etc.), ask the user before rerunning.
© UKGovernmentBEIS, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in .claude/skills/scientific-debug of UKGovernmentBEIS/vllm-lens.
Open the folder on GitHubat commit 3d11fb0
Compare with similar skills
Scientific Debug 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 |
|---|---|---|---|---|---|---|
| Scientific Debug this skillUKGovernmentBEIS/vllm-lens | 132 | — | ~556 | Automated safety check: Pass | MIT | |
| Aoti Debugpytorch/pytorch | 104k | 1 repos | ~1.7k | Automated safety check: Pass | Custom licence | |
| 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 | |
| Release DigestKiln-AI/Kiln | 5.2k | — | ~2.7k | Automated safety check: Pass | Custom licence |
Similar skills
Debug Distributed Hang
sgl-project/sglang
Debug hanging issues in SGLang distributed inference (TP/PP/DP/EP).
- 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.
Release Digest
Kiln-AI/Kiln
Post a "what's changed since the last release" recap to the release Slack channel for final QA.
Veomni Debug
ByteDance-Seed/VeOmni
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…
Related
Categories
Questions about Scientific Debug
What does Scientific Debug do?
A skill your agent uses whenever the user asks to debug something, fix a bug, troubleshoot an issue, or mentions "scientific debugging". Scientific Debug is an agent skill from UKGovernmentBEIS/vllm-lens. Use this skill whenever the user asks to debug something, fix a bug, troubleshoot an issue, or mentions "scientific debugging".
When should I use Scientific Debug?
Scientific Debug fits situations like: the user asks to debug something; troubleshoot an issue; mentions scientific debugging.
How do I install Scientific Debug in Claude Code?
Run `npx skills add UKGovernmentBEIS/vllm-lens --skill scientific-debug -a claude-code`. Or copy the skill folder (.claude/skills/scientific-debug in UKGovernmentBEIS/vllm-lens) into .claude/skills/scientific-debug in your project. Claude Code loads it when a task matches its description.
How do I install Scientific Debug in Codex?
Run `npx skills add UKGovernmentBEIS/vllm-lens --skill scientific-debug -a codex`. Or copy the skill folder (.claude/skills/scientific-debug in UKGovernmentBEIS/vllm-lens) into .agents/skills/scientific-debug in your project. Codex loads it when a task matches its description.
Can I use Scientific Debug 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 UKGovernmentBEIS/vllm-lens --skill scientific-debug -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/scientific-debug, .gemini/skills/scientific-debug, .github/skills/scientific-debug and .opencode/skills/scientific-debug in your project.
What does Scientific Debug need to run?
SKILL.md names no scripts, command-line tools or credentials: Scientific Debug is instructions for the agent only.
Does Scientific Debug 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 Scientific Debug 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 Scientific Debug use?
Scientific Debug is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Scientific Debug use?
About 556 tokens (SKILL.md is roughly 2.2k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Scientific Debug?
Skills that share tags, products or a category with Scientific Debug: Aoti Debug (pytorch/pytorch, 104k stars), Debug Distributed Hang (sgl-project/sglang, 37k stars), CUTLASS FMHA Incremental Rebuild (microsoft/onnxruntime, 22k stars) and The Art of Debugging (stas00/the-art-of-debugging, 1.7k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Scientific Debug?
UKGovernmentBEIS (a GitHub organization) maintains it in UKGovernmentBEIS/vllm-lens, which has 132 GitHub stars. The repository was last updated on October 2, 2026.
Source: UKGovernmentBEIS/vllm-lens on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.