Verified:GitHub

read-paper

Generate a structured paper reading workspace from an arXiv PDF and TeX source.

1|Updated Mar 17, 2026

Search Tags:#arxiv #figure extraction #structured workflow #deep reading #note generation #image prioritization #pdf to png

Ready to run

npx skills add https://github.com/bobjason2020/read-paper --skill read-paper-bobjason2020

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: read-paper
Source: https://github.com/bobjason2020/read-paper/tree/main
Command: npx skills add https://github.com/bobjason2020/read-paper --skill read-paper-bobjason2020

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

许多论文阅读流程需要反复手动下载、提取图表、整理阅读笔记与“图文对应”的结构化内容，导致理解效率低且容易遗漏关键信息。

Core Features & Use Cases

Two-stage workflow（脚本机械化 + AI深度分析）：先自动创建论文工作区（下载PDF/TeX、提取图片并生成 images.md 与笔记模板），再由AI进行图片视觉判断、精读并填充完整精读笔记。
图表优先级与图文并茂插入：通过 AI 视觉逐张判断高/中/低优先级，并强制在笔记的对应章节插入，而不是把所有图片堆在末尾。
结构化笔记模板与任务追踪：生成可直接填充的精读笔记模板，并要求在阶段二用 TaskCreate/TaskUpdate 追踪关键步骤，降低跳步风险。

Quick Start

Tell your AI assistant: 精读论文 https://arxiv.org/abs/2401.12345，并要求先运行阶段一生成 images.md 与笔记模板，再按任务列表先逐张视觉判断图片优先级、调用 process_selected_images.py 处理选中图片，最后阅读PDF/TeX并填充完整笔记。

Frequently Asked Questions about read-paper

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I extract figures from an arXiv PDF for deep reading notes?▼

This workflow downloads the arXiv PDF and TeX source, converts figures to PNG images, and generates an images.md file for deep reading notes. It automates figure extraction to prevent manual errors and prepare a structured workspace for visual analysis.

What is a structured workflow for arXiv paper deep reading and note generation?▼

A structured arXiv paper deep reading workflow uses two stages: scripts mechanize PDF and TeX source downloading plus figure extraction, while AI performs visual prioritization and fills a structured note template. This ensures consistent figure-to-section mapping across architecture, methods, and experiments.

How do I prioritize figures visually when reading a research paper?▼

The AI stage evaluates each extracted image to assign high, medium, or low priority for visually prioritizing figures when reading a research paper. It then mandates inserting these prioritized images directly into their corresponding note sections instead of grouping them at the end.

Can I use TeX source to improve PDF to PNG figure extraction?▼

Yes, you can use TeX source to improve PDF to PNG figure extraction. By downloading both the arXiv PDF and TeX source, the scripts component accurately extracts figures and generates an images.md file, ensuring higher quality image handling than PDF-only parsing.

Do I need TaskCreate tracking for arXiv deep reading workflows?▼

Yes, you need TaskCreate tracking for arXiv deep reading workflows. The second stage enforces mandatory TaskCreate and TaskUpdate tracking to monitor critical steps like visual prioritization and script execution, reducing the risk of skipping essential reading tasks.