ARIS (Auto-claude-code-research-in-sleep)

Other

Autonomous ML research skills

Use it when you want Claude to rework a research paper overnight with Codex as strict reviewer

A collection of 80+ Claude Code skills that run research tasks like literature review, experiments, paper writing and cross-model review loops.

Added to the shelf 16 Oct at 14:05

From the makersREADME, read 16 Oct

ARIS (Auto-Research-In-Sleep) is a set of lightweight Markdown-only skills for autonomous ML research, covering cross-model review loops, idea discovery and experiment automation. It is a methodology rather than a platform, with no framework and no lock-in. It works with Claude Code, Codex CLI, Cursor and other LLM agents, and its reviewer is an independent model, Codex MCP by default.

How creators use it1 take

@buildwithbyronInstagram, 16 Oct
Scores, experiments on and rewrites a research paper overnight using Claude plus Codex review.
1. Give it your paper draft before bed and run the auto review loop so Claude fixes what is wrong and repeats
  /auto-review-loop

Kinds of work