mean-reviewer

What problem does it solve?

Academic peer review is often superficial, inconsistent, or hostile. This skill demonstrates how easily an LLM can generate a convincingly authoritative, structurally devastating review that exposes the vulnerability of the current system. It helps researchers understand what a bad review looks like and stress-test their papers against adversarial critique before submission.

Core Features & Use Cases

• Adversarial Review Generation: Produces a 20+ point review mixing legitimate concerns with inflated, unfalsifiable objections, evaluation framework attacks, and fixed low scores.
• Rebuttal Simulation: When given an author response, generates a post-rebuttal reply that dismisses new experiments, weaponizes concessions, and refuses to raise the score.
• Educational Demonstration: Uses a real NeurIPS 2025 oral paper to show how a maximally destructive reviewer can sink excellent work, highlighting systemic vulnerabilities in academic peer review.

Quick Start

Use the mean-reviewer skill to generate a devastating peer review of the paper text or abstract you provide, and follow up with an author rebuttal to simulate the full review cycle.