# Academic Paper Review

> Use when: review an academic paper with a structured review form and multiple reviewer personas.

- Skill: `kimtth/academic-paper-review` (Agent Skill)
- Install (CLI): `npx skillmds@latest add kimtth/academic-paper-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kimtth/academic-paper-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: kimtth (https://skillmd.com/u/kimtth)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kimtth/academic-paper-review

---


Goal: a fair, specific review that separates contribution from flaws.

Use for:
- reviewing a paper before submission or as a referee
- giving structured feedback on a draft
- assessing whether claims are supported by evidence

Workflow:
1. Load the paper; extract its text and section structure.
2. Review from three personas: harsh-but-fair, harsh-critical, open-minded.
3. For each persona, score with a structured review form (e.g. NeurIPS).
4. Reflect and refine each review for accuracy and soundness.
5. Aggregate: average scores, weighting overall and contribution highest.
6. Write a meta-review: consensus strengths, weaknesses, and a decision.

Assess:
- novelty relative to prior work
- soundness of methodology and analysis
- adequacy of baselines, ablations, and evaluation
- honesty about limitations and threats to validity

Rules:
- tie every critique to a specific passage or result
- find consensus across personas before deciding
- separate major issues from minor ones; suggest fixes
- judge the evidence, not the abstract's claims

Rules:
- tie every critique to a specific passage or result
- distinguish "wrong" from "not my preference"
- check claims against the actual evidence, not the abstract
- be rigorous and fair; acknowledge real strengths