skills.rest
lhwcvlhwcvCommunity·2 Agent Skills Included

tinyKaggleClaw

Multi-agent runtime for continuous Kaggle and ML research

Runs a standing three-role team (leader, researcher, trainer) that iterates on Kaggle and ML tasks from a single kickoff request. Handles EDA, baseline coding, training queue submission, result summaries, and trend charts without repeated manual prompting. Includes a forum-style runtime board for watching agent discussions and injecting ideas, plus a separate training queue board for job tracking.
60|13
↗
npx skills add lhwcv/tinyKaggleClaw --all -g -y
Available:

Defines the shared operating contract for the leader, researcher, and trainer agents, covering inbox delegation, EDA-first workflow, versioned baselines, and result documentation rules.

All Skills in This Repository (2)

Pure Emerald Level Indicators
📦 In Repo
lhwcvlhwcv

kaggle-kernel-discussion-birdclef2026

Pull and summarize BirdCLEF 2026 Kaggle discussions and notebooks into structured files.

Community
Advanced
📦 In Repo
lhwcvlhwcv

Runtime Communications

Coordinate and monitor ML research runtime tasks across leader, researcher, and trainer agents.

Community
Intermediate

Frequently Asked Questions

FAQPage Schema
How to install tinyKaggleClaw?▼

Run `npx skills add lhwcv/tinyKaggleClaw --all -g -y` in your terminal to install all skills in this suite globally.

How to automate Kaggle competition research?▼

Start the runtime and send one task to the leader agent; the team then performs EDA, builds versioned baselines, submits training jobs, and writes result summaries on its own.

What do the leader, researcher, and trainer agents do?▼

The leader orchestrates and delegates work, the researcher handles EDA and experiment implementation, and the trainer submits jobs to the training queue and writes result reports.

Can I watch and guide the agents while they work?▼

Yes. The forum-style runtime board at localhost:8090 lets you watch agent conversations and inject ideas or corrections without taking over the workflow.

Does tinyKaggleClaw track experiment progress over time?▼

Yes. The trainer agent writes docs for every baseline version and generates trend charts comparing the top experiments across versions.

Related Repositories in Data & Analytics

View All in Data & Analytics→
PaddlePaddlePaddlePaddle
87.3k

PaddleOCR

Extract text, tables, and formulas from PDFs and images

Official📦 2 Skills
D4VinciD4Vinci
73.3k

Scrapling

Scrape any website and bypass anti-bot protection with AI

Community📦 1 Skills
mvanhornmvanhorn
60.8k

last30days-skill

Research any topic across Reddit, X, YouTube, and the web

Community📦 2 Skills