# Auto-claude-code-research-in-sleep

> Auto-Research-in-Sleep (ARIS): Autonomous machine learning and academic research workflow. Supports overnight multi-agent idea discovery, code execution, adversarial cross-model review loops, narrative report synthesis, and LaTeX submission rebuttal.

- Skill: `marvin-lijunzan/auto-claude-code-research-in-sleep` (Agent Skill)
- Install (CLI): `npx skillmds@latest add marvin-lijunzan/auto-claude-code-research-in-sleep`
- Raw SKILL.md: https://api.skillmd.com/api/skills/marvin-lijunzan/auto-claude-code-research-in-sleep/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: Marvin-LiJunZan (https://skillmd.com/u/marvin-lijunzan)
- Updated: 2026-09-21
- Page: https://skillmd.com/skills/marvin-lijunzan/auto-claude-code-research-in-sleep

---


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

Autonomous overnight machine learning research and adversarial multi-model auto-review loops for Claude Code, Codex, and Antigravity.

---

## Capabilities & Architecture

1. **Idea Discovery & Experiment Bridge**:
   - Converts high-level research questions into verifiable numerical hypotheses and PyTorch experiment runs.
2. **Cross-Model Adversarial Auto-Review Loops**:
   - Multi-agent peer review pitting Claude, Gemini, and GPT against each other to ruthlessly identify experimental gaps, baseline flaws, and overclaimed conclusions.
3. **Manuscript & Rebuttal Generation**:
   - Full publication drafting, LaTeX packaging, reviewer point-by-point rebuttal drafting.

The canonical implementation and sub-modules are located in `skills/auto-research-in-sleep/`.