🧪
QualityPython

ML Debugging

by Leeroo-AI

ML Debugging is a Quality skill for Claude Code, published by Leeroo-AI in superml.

194 stars18 forkson Leeroo-AI/supermlAdded 2026/09/05+1% in starsRepository updated 2026/03/17
View source on GitHubDownload repo ZIP
aiclaude-codecodexcoding-agent-skillscoding-agentscursor-aillmmachine-learningmlmulti-agentpluginskill
Install in seconds
Install ML Debugging
Copy ML Debugging into your Claude Code skills folder. Run the command in your terminal, or review the source on GitHub before installing.
terminal
npx degit https://github.com/Leeroo-AI/superml/tree/main/skills/ml-debug ~/.claude/skills/ml-debug

Requires Node.js. Downloads this skill only — not the rest of the repository — into your Claude Code skills folder.

Without Node.js

git clone https://github.com/Leeroo-AI/superml.git

Clones the whole repository, then copy the skill’s own directory into your skills folder yourself.

In this catalog

Source file
skills/ml-debug/SKILL.md in Leeroo-AI/superml
Installs to
~/.claude/skills/ml-debug
Collection
One of 7 skills cataloged from this repository
Category
Quality — 1662 skills

What ML Debugging does

ML Debugging systematically diagnoses ML and AI training failures such as OOMs, NaNs, divergence, crashes, slow throughput, wrong outputs, and dependency conflicts. Use it when an experiment or model run breaks and you need root-cause analysis before applying fixes.

ML Debugging is cataloged under Quality on DirSkills. ML Debugging comes from a repository tagged ai, claude-code, codex, coding-agent-skills and coding-agents.

Documentation

README

ML Debugging

Systematically diagnose ML failures using framework-specific knowledge, not guesswork.

Grounding

Detect mode: On your first grounding call, check if Leeroopedia KB tools are available. If they return results, use KB mode. If unavailable or auth fails, use Web mode.

HARD RULE: You MUST ground before writing analysis. If KB fails, you MUST WebFetch at least 2 URLs before writing ANY diagnosis. Writing from memory without fetching is the #1 failure mode of this skill — it produces zero-citation responses that score 0/3 on grounding. "I know X well" is NOT a substitute for fetching documentation.

KB mode: Call diagnose_failure → query_hyperparameter_priors → search_knowledge. Cite as [PageID].

This is the opening of the README. Read the full README on GitHub.

Frequently asked about ML Debugging

  • What else does Leeroo-AI publish alongside ML Debugging?

    ML Debugging is one of 7 skills that DirSkills catalogs from Leeroo-AI/superml, the repository it ships in. Its siblings there include ML Experiment Journal, ML Iteration and ML Planning. Each one is a separate skill with its own page in this directory, installs the same way ML Debugging does, and is maintained by Leeroo-AI in that same repository. The rest of the collection is listed on the Leeroo-AI/superml page.

  • How does ML Debugging compare to other Quality skills?

    ML Debugging ranks #1549 by stars among the 1662 Quality skills in this catalog. The most-starred ones next to it are Benchmark, Benchmark Optimization Loop and API Design Patterns. DirSkills ranks by the star count of the repository each skill ships in, so that order reflects how popular those repositories are rather than any review of ML Debugging against them. Open each page to compare what they document and how they install.

More from Leeroo-AI/superml

ML Debugging is one of 7 skills cataloged on DirSkills from Leeroo-AI/superml.

See all 7 skills →
🧪
43m ago

ML Experiment Journal

ML Experiment Journal keeps a persistent log of experiment hypotheses, results, and lessons across sessions. Use it before, during, and after ML runs to track what changed and what was learned.
Data
19418
🧪
43m ago

ML Iteration

ML Iteration generates grounded, ranked next steps when an ML experiment is stuck or underperforming. Use it after initial runs to compare alternatives, review what was tried, and decide what to change next.
AI Engineering
19418
🧠
43m ago

ML Planning

ML Planning turns an ML goal into a grounded, step-by-step implementation plan. Use it when you need an architecture, pipeline, or build plan for a machine learning project.
AI Engineering
19418
🔬
43m ago

ML Research

ML Research helps compare ML and AI approaches using verified framework documentation and model references. Use it when you need to understand how a method works or survey framework capabilities.
AI Engineering
19418
🔎
43m ago

ML Verification

ML Verification checks code, configs, and math before training or deployment to catch mistakes early. It is used to validate framework behavior and flag issues before expensive runs.
Quality
19418
📚
43m ago

Using Leeroopedia

Using Leeroopedia establishes the lookup-first workflow for ML and AI conversations. It tells the assistant to use Leeroopedia or web sources, cite claims, and check for config or API mistakes before answering.
AI Engineering
19418

More Quality skills

ML Debugging is one of 1662 Quality skills cataloged on DirSkills, ranked #1549 by stars among them.

Browse all Quality skills →
⏱️
3w ago

Benchmark

Benchmark measures performance baselines, detects regressions before and after PRs, and compares stack alternatives for pages, APIs, and builds.
Quality
239.8K36.4K
⏱️
3w ago

Benchmark Optimization Loop

Benchmark Optimization Loop turns vague performance goals into a bounded, measured optimization process. Use it when asked to make something faster, benchmark latency/throughput/cost, or choose the best implementation by repeated measured tests.
Quality
239.8K36.4K
🔌
3w ago

API Design Patterns

API Design Patterns provides conventions and best practices for designing consistent REST APIs. Use it when creating endpoints, reviewing contracts, handling pagination, error responses, status codes, filtering, or versioning.
Quality
239.8K36.4K