academic-paper-review

academic-paper-review produces structured, peer-review-quality analyses of academic papers, preprints, and scientific publications. It addresses the problem of turning a raw paper (an arXiv link, DOI, journal URL, or uploaded PDF) into a rigorous, balanced assessment that mirrors the standards of top-tier venues such as NeurIPS, ICML, ACL, Nature, and IEEE, with every judgment grounded in evidence from the paper itself.

The methodology runs in three phases. Comprehension extracts paper metadata, performs a systematic deep read of abstract through conclusion, and lists the paper's main claims with supporting evidence and a strength rating. Critical analysis uses web search and web-fetch to position the work in the current research landscape, then scores methodology across soundness, novelty, reproducibility, experimental design, statistical rigor, and scalability, and assigns a contribution significance level from landmark down to below-threshold. Synthesis assembles a structured review with an executive summary, contribution list, and titled strengths and weaknesses, each anchored to specific sections, figures, or tables. It can output both a detailed review and a concise summary, and supports papers across any scientific domain.

The skill targets researchers, reviewers, students, and program committee members who need to review, critique, or summarize research, or prepare a peer review for a conference or journal submission. It is a read-and-analyze methodology whose only external actions are web search and fetching public papers for literature context; it does not modify systems or handle credentials.