ARIS — Auto Research in Sleep
Autonomous ML research loops that run cross-model review, idea discovery, and experiment automation while you sleep. Designed for overnight batch research.
ARIS (Auto Research in Sleep) is a skill pack for Claude Code that runs autonomous ML research pipelines — literature review, idea generation, experiment execution, and paper writing — with cross-model review, where a second model (e.g. GPT via the Codex MCP) audits Claude's work. Installed by cloning the repo and symlinking its skills into your project; adaptations exist for Codex CLI, Cursor, and other agent hosts.
- Queue up research experiments to run overnight while you sleep
- Discover new approaches to a problem through automated exploration
- Get cross-checked research results reviewed by multiple AI models
Autonomous ML research loops that run cross-model review, idea discovery, and experiment automation while you sleep. Designed for overnight batch research. ~1.2K stars.
ML researchers and data scientists who want to queue up research tasks for Claude to execute autonomously during off-hours.
https://github.com/wanshuiyin/Auto-claude-code-research-in-s...
By wanshuiyin
How to Get It
git clone https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep.git bash Auto-claude-code-research-in-sleep/tools/install_aris.sh ~/your-project
After installing, paste this into Claude:
Help me queue up research experiments to run overnight while me sleep
Trust Signals Reviewed
Community Pulse Active
Discussed on Reddit
- 25 Claude Code Tips from 11 Months of Intense Use — Reddit · 538 pts
- Stop wasting time: These 12 FREE AI tools can literally run your work while you — Reddit · 82 pts
- Claude Code's Auto Dream feature explained — here's the full technical breakdown — Reddit · 30 pts
3 mentions across 1 sources
Reviewer notes
Reviewed review. These are observations, not a security certification.
Niche ML research skills. Star count may be inflated vs actual adoption. Legitimate Claude Code skills.
Things to check
- Requires careful prompt engineering and monitoring to avoid divergent experiment directions; token costs scale with experiment depth and may accumulate silently in background runs. No built-in safety guardrails for resource-intensive loops.
How to evaluate tools before deploying →
Data shown here comes from public APIs and automated scanning. Reviewer notes reflect one person's experience. This is not a security certification or legal recommendation. Always evaluate tools according to your own organization's policies.
Evaluation
Niche ML research skills. Star count may be inflated vs actual adoption. Legitimate Claude Code skills.