Back to Details

self-review

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Self-Review

自我评审

Review an academic paper using a structured review form with multiple reviewer personas.
使用带多种评审人员角色的结构化评审表格评审学术论文。

Input

输入

  • $ARGUMENTS
    — Path to PDF file or 
    .tex
    file
  • $ARGUMENTS
    — PDF文件或
    .tex
    文件的路径

Scripts

脚本

Extract text from PDF

从PDF提取文本

bash
python ~/.claude/skills/self-review/scripts/extract_pdf_text.py paper.pdf --output paper_text.txt
python ~/.claude/skills/self-review/scripts/extract_pdf_text.py paper.pdf --format markdown
Tries pymupdf4llm (best) → pymupdf → pypdf. Install: 
pip install pymupdf4llm pymupdf pypdf
bash
python ~/.claude/skills/self-review/scripts/extract_pdf_text.py paper.pdf --output paper_text.txt
python ~/.claude/skills/self-review/scripts/extract_pdf_text.py paper.pdf --format markdown
优先尝试pymupdf4llm（效果最佳）→ pymupdf → pypdf。安装命令：
pip install pymupdf4llm pymupdf pypdf

Parse PDF into structured sections

将PDF解析为结构化章节

bash
python ~/.claude/skills/self-review/scripts/parse_pdf_sections.py \
  --pdf paper.pdf --output sections.json
Extracts title (via font size), section headings, and section text. Requires: 
pip install pymupdf
Key flags: 
--format text
, 
--verbose
bash
python ~/.claude/skills/self-review/scripts/parse_pdf_sections.py \
  --pdf paper.pdf --output sections.json
提取标题（通过字体大小识别）、章节标题和章节文本。依赖：
pip install pymupdf
关键参数：
--format text
, 
--verbose

Workflow

工作流程

Step 1: Load Paper

步骤1：加载论文

  • If PDF: use 
    extract_pdf_text.py
    to extract text
  • If 
    .tex
    : read the LaTeX source directly
  • 若为PDF文件：使用
    extract_pdf_text.py
    提取文本
  • 若为
    .tex
    文件：直接读取LaTeX源码

Step 2: Three-Persona Review

步骤2：三角色评审

Run three independent reviews using different personas (from 
references/review-form.md
):
  1. Harsh but fair reviewer: Expects good experiments that lead to insights
  2. Harsh and critical reviewer: Looking for impactful ideas in the field
  3. Open-minded reviewer: Looking for novel ideas not proposed before
For each persona, generate a review following the NeurIPS review JSON format in 
references/review-form.md
.
使用三种不同的评审人员角色（定义于
references/review-form.md
）执行三次独立评审：
  1. 严格但公正的评审员：期望实验能得出有价值的洞见
  2. 严苛且批判性的评审员：关注领域内有影响力的创新想法
  3. 思想开放的评审员：寻找此前未被提出的新颖想法
针对每个角色，按照
references/review-form.md
中的NeurIPS评审JSON格式生成评审意见。

Step 3: Reflection Refinement (up to 3 rounds per reviewer)

步骤3：反思优化（每位评审员最多3轮）

After each review, apply the reflection prompt: re-evaluate accuracy and soundness, refine if needed. Stop when "I am done".
每次评审后，应用反思提示词：重新评估评审的准确性与合理性，如有需要则优化。当评审员回复"完成"时停止。

Step 4: Aggregate

步骤4：结果汇总

  • Combine all three reviews
  • Average numerical scores (round to nearest integer)
  • Synthesize a meta-review finding consensus
  • Weight scores using AgentLaboratory weights: Overall (1.0), Contribution (0.4), Presentation (0.2), others (0.1 each)
  • 合并三份评审意见
  • 对数值评分取平均值（四舍五入至整数）
  • 综合生成一份达成共识的元评审意见
  • 使用AgentLaboratory权重对评分进行加权：总体评分（1.0）、贡献度（0.4）、表述质量（0.2），其余维度各占0.1

Step 5: Actionable Report

步骤5：可落地的评审报告

Output format:
undefined
输出格式：
undefined

Review Summary

评审摘要

  • Overall Score: X/10 (Weighted: Y/10)
  • Decision: Accept / Reject
  • Confidence: Z/5
  • 总体评分：X/10（加权后：Y/10）
  • 评审结论：录用 / 拒绝
  • 置信度：Z/5

Strengths (consensus across reviewers)

优势（评审员共识）

  1. ...
  2. ...
  1. ...
  2. ...

Weaknesses (consensus across reviewers)

不足（评审员共识）

  1. ...
  2. ...
  1. ...
  2. ...

Questions for Authors

给作者的问题

  1. ...
  1. ...

Specific Suggestions for Improvement

具体改进建议

  1. [章节X，第Y页]：...
  2. [章节Z，第W页]：...

Score Breakdown

评分明细

DimensionR1R2R3Avg
Overall............
Contribution............
...............
undefined
维度评审1评审2评审3平均值
总体............
贡献度............
...............
undefined

References

参考资源

  • NeurIPS review form, scoring weights, personas, reflection prompts: 
    ~/.claude/skills/self-review/references/review-form.md
  • PDF text extraction: 
    ~/.claude/skills/self-review/scripts/extract_pdf_text.py
  • NeurIPS评审表格、评分权重、评审角色、反思提示词：
    ~/.claude/skills/self-review/references/review-form.md
  • PDF文本提取脚本：
    ~/.claude/skills/self-review/scripts/extract_pdf_text.py

Missing Sections Check

缺失章节检查

You MUST verify that all required sections are present: Abstract, Introduction, Methods/Approach, Experiments/Results, Discussion/Conclusion. Reduce scores if any are missing.
必须验证所有必填章节是否存在：摘要、引言、方法/研究途径、实验/结果、讨论/结论。若有缺失，需降低评分。

Related Skills

相关技能

  • Upstream: paper-compilation
  • Downstream: paper-revision, rebuttal-writing
  • See also: slide-generation
  • 上游技能：论文汇编
  • 下游技能：论文修订、反驳信撰写
  • 其他相关：幻灯片生成