ARIS (Auto-Research-In-Sleep)
About this tool
Around 80 Markdown-only skills for ML research. A cross-model review loop uses one model to work and another model family to review. /research-pipeline connects literature searches, topic selection, experiments, automated reviews and writing. Additional skills cover rebuttals, slides, posters and citation checks in hosts such as Claude Code, Codex and Cursor.
- Feature 1: /research-pipeline connects research, topic selection, experiments, review and writing.
- Feature 2: idea-discovery explores topics; /auto-review-loop runs cross-model adversarial review.
- Feature 3: experiment-bridge and experiment-queue automate experiments.
- Feature 4: rebuttal, paper-slides, paper-poster-html and citation-crosscheck prepare submission materials.
- Feature 5: Markdown-only skills without framework dependencies.
Resource information
- Maintainer
- Independent developer wanshuiyin (Yang Ruofeng, whose profile lists Shanghai Jiao Tong University) and community contributors.
- Type
- Skill collection
- Category
- Research skill collections
- Runs on
- Agent Skills
- Who it's for
- AI and ML research teams running experiments, reviewing and revising papers, or preparing rebuttals, slides and posters.
- License at time of listing
- MIT
- Source
- GitHub
