Paper Reviewer

Academic Writing

Simulates a notoriously strict senior conference reviewer to help authors identify fatal flaws before submission.

Overview

This skill simulates a harsh but constructive top-conference reviewer. It approaches paper review with a rejection mindset, focusing on identifying fatal issues that could lead to direct rejection. The output includes a structured review report (summary, strengths, critical weaknesses, rating) and strategic revision advice, all in Chinese for maximum accessibility to the target audience.

When to Use

• Pre-submission review before sending to a top conference (ICML, NeurIPS, ICLR, CVPR, ACL, AAAI)
• Identifying potential rejection reasons that you might have missed
• Getting a reality check on claimed contributions vs actual experimental evidence
• Stress-testing your paper's logical consistency and experimental rigor

Key Features

• Rejection-first mindset: defaults to looking for reasons to reject, only changes if highlights are compelling
• Specific critiques: says "lacks robustness validation on ImageNet" instead of vague "experiments insufficient"
• 3-dimensional review: evaluates originality, rigor (math derivations, fair baselines, ablations), and consistency (claims vs evidence)
• Structured output: Review Report (Summary, Strengths, Critical Weaknesses 3-5, Rating 1-10) + Strategic Advice (root cause analysis + action guide)
• Chinese output: both the review and revision advice are in Chinese for direct usability

Example Prompts

Review this paper from a strict ICML reviewer perspective.

Simulate a NeurIPS reviewer and identify fatal flaws in my paper.

I plan to submit to ICLR 2026, please critique this paper harshly.

Source

• Skill folder: skills/paper-reviewer/
• Origin: awesome-ai-research-writing
• Standard: agentskills.io