What it does
A set of Markdown-only skills for running machine learning research with an AI agent. It covers idea discovery, experiment automation and review loops where one model checks the work of another. It needs no framework and works with Claude Code, Codex or other LLM agents.
Use cases
- 01Explore research ideas before committing to one
- 02Automate a series of machine learning experiments
- 03Have a second model review a draft or result