Leeroo-AILeeroo-AIOfficialยท7 Agent Skills Included

superml

Grounded ML planning, debugging, verification, and experiment tracking

Guides machine learning workflows from planning and config verification to debugging OOM, NaN, and divergence failures by root cause. Eliminates wasted GPU hours, stale training-data guesses, and repeated failed experiments through documentation-grounded answers with citations. Connects to a 27k-page ML knowledge base and keeps a persistent experiment journal so results and lessons carry across sessions.
npx skills add Leeroo-AI/superml --all -g -y

All Skills in This Repository (7)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install SuperML?โ–ผ

Run `npx skills add Leeroo-AI/superml --all -g -y` in your terminal to install all seven ML workflow skills globally.

What does SuperML do for ML engineers?โ–ผ

It helps you plan training runs, verify configs before burning GPU hours, debug failures like OOM and NaN by root cause, and get ranked next steps when metrics plateau. Every answer is grounded in real framework documentation with citations.

Does SuperML work with Claude Code and Cursor?โ–ผ

Yes. It works with Claude Code, Cursor, Codex, OpenCode, and Gemini CLI, and a session hook loads the skills automatically with zero per-conversation setup.

Do I need an API key to use SuperML?โ–ผ

No. Without a key, the skills fall back to web search and official documentation for grounding. With a free Leeroopedia key, your agent gets faster access to 27k+ curated ML framework pages.

How does SuperML prevent repeated failed experiments?โ–ผ

The ml-experiment skill maintains a persistent journal of hypotheses, results, and lessons across sessions, so your agent checks what already failed before proposing new runs.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’