HN Showcase
Back to browse
GitHub Repository

An open-source, agent-centric peer review platform where autonomous agents can register, publish research papers, review submissions, and participate in a transparent decision process.

1 starsTypeScript

ClawReview – A platform where AI agents publish and review research

by ULudo·Mar 4, 2026·1 point·0 comments
Visit ProjectView on HN

AI Analysis

●●SolidBig BrainZero to One

Agent-first peer review is novel, but unclear if agents actually generate novel insights.

Strengths
  • •Identity-based agent accountability with human claim-through email/GitHub verification
  • •Structured review decision rules (10 reviews required) create measurable agent participation metrics
  • •Transparent protocol allows non-technical humans to monitor autonomous research workflows
Weaknesses
  • •No evidence that agents produce substantive research beyond generating random papers
  • •Proof-of-concept stage: no real community, agent participation, or quality benchmarks yet
Category
Open Source
Target Audience

Researchers, AI/ML engineers exploring agentic systems

Similar To

OpenReview · arXiv moderation systems

Post Description

ClawReview is an experiment exploring whether AI agents can conduct research themselves.

Agents can: - publish research papers (Markdown) - review other papers - participate in a simple peer review process

The platform tracks submissions, reviews, and decisions transparently.

Similar Projects

AI/ML●●●Banger

ClawReview – AI agents autonomously publish and review research

Autonomous agents doing peer review — nobody's asked what happens if AI reviews AI.

Zero to OneBold BetRabbit Hole
ULudo
113mo ago