Automated academic paper review system for Claude Code. Integrates PDF parsing (marker-pdf + PyMuPDF), reference verification (CrossRef + DBLP + Semantic Scholar), SOTA comparison, and competitive paper analysis within a structured multi-phase review pipeline. Features venue-adaptive review strategies for 15+ top CS conferences (CVPR, NeurIPS, ICML, ICLR, ECCV, ICCV, ACL, AAAI, KDD, MICCAI, etc.), powered by distilled methodologies from 10 senior conference reviewers. Includes 8-dimension scoring, statistical rigor checks, automated quick assessment, rebuttal assistance, and parallel agent-based review architecture. Components: 3 commands (/review-paper, /pre-screen, /rebuttal), 4 specialized agents, 1 skill with 5 reference files, 2 MCP servers (11 tools total), automated dependency checking, and a quick assessment script.