Auto Paper Reading

Q: What does the Auto Paper Reading skill do?

当用户要求"讲解 / 精读 / 解读 / 复盘一篇论文"，或提供 .pdf/.docx/.doc 论文文件， 或只给出论文题目希望自动搜索下载，并输出图文结合的讲解 Word 文档时使用。 工作流：搜索下载论文 → 解析裁剪关键图 → 视觉分析选图 → 查找 baseline 与领域问题 → 撰写摘要、符号表与图文结合的五部分讲解 → 生成 docx → 校验排版 → 对话中讲解。 触发词：讲论文、读论文、论文讲解、paper reading、帮我读懂这篇论文 等。 Also use for reading or explaining a research paper and producing an illustrated Word report in Chinese or English, from a title or a PDF/DOCX/DOC file.

当用户要求"讲解 / 精读 / 解读 / 复盘一篇论文"，或提供 .pdf/.docx/.doc 论文文件，或只给出论文题目希望自动搜索下载，并输出图文结合的讲解 Word 文档时使用。工作流：搜索下载论文 → 解析裁剪关键图 → 视觉分析选图 → 查找 baseline 与领域问题 → 撰写摘要、符号表与图文结合的五部分讲解 → 生成 docx → 校验排版 → 对话中讲解。触发词：讲论文、读论文、论文讲解、paper reading、帮我读懂这篇论文等。 Also use for reading or explaining a research paper and producing an illustrated Word report in Chinese or English, from a title or a PDF/DOCX/DOC file.

5 starsResearch & Science

Install

skills CLI (Claude Code, Cursor, Codex and more)

npx skills add berglingtim/auto_paper_reading --skill auto-paper-reading

Manual install

git clone --depth 1 https://github.com/berglingtim/auto_paper_reading.git /tmp/auto_paper_reading
mkdir -p ~/.claude/skills
cp -r /tmp/auto_paper_reading/. ~/.claude/skills/auto-paper-reading

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# auto_paper_reading 论文讲解工作流

[简体中文](SKILL.md) | [English workflow](SKILL.en.md)

把一篇论文（`.pdf` / `.docx` / `.doc`）变成一份**图文结合的中文或英文讲解 docx**，
并在对话里把论文讲给用户听（验收标准：用户满意）。
本工具**不调用任何 LLM API**：搜索下载、看图选图、撰写讲解全部由当前 Agent
（需具备联网检索与读图能力）亲自完成；脚本只负责解析、裁剪与排版。

## 语言选择（默认中文）

- 优先遵循用户明确指定的报告语言；未指定时默认中文，不因论文原文是英文就切换。
- 在 `content.json` 顶层写 `"language": "zh"` 或 `"language": "en"`。英文工作流见
  [SKILL.en.md](SKILL.en.md)；两种语言使用同一套证据、baseline、领域检索与排版质量标准。
- Agent 用目标语言撰写标题、摘要、符号解释、章节名、正文、图注、表格内容与对话讲解。
  原文引用、论文名、算法名可保留原语言；引用论文原图不改图中文字。排版器只切换生成标签，
  不自动翻译正文，不增加任何 LLM API 调用。
- DOCX 阶段优先级为 `--language` > `content.json.language` > `zh`。prep 阶段的语言选项
  只控制命令行提示；撰写内容时仍要设置 `language`。只支持 `zh` / `en`，无效值报错。
- 双版本分别写到不同工作目录，例如 `--out output/zh` 与 `--out output/en`；输入为
  `papers/paper.pdf` 时，对应 `output/zh/paper/content.json` 和 `output/en/paper/content.json`。
  所有命令都会在 `--out` 后追加输入文件名（不含扩展名），以实际输出日志为准。
- 中文一级标题为“一、二、三……”，英文为“1、2、3……”；二/三级均用 `1.1 / 1.1.1`。
  英文自动标签为 Abstract / Notation，公式编号为 `(1)`，文件名前缀为 `Explanation_`；
  中文保持摘要 / 符号表、`（1）` 与 `讲解_`。报告结构、黑白三线表及双向居中标准一致。
- Word 输入转换与预览默认使用免费开源的 LibreOffice；WPS 仅为显式可选后端。安装与实际验证平台见 [README](README.md#pdf-转换后端)。
  缺少转换器时可生成 DOCX，但应如实报告视觉检查未完成；双语和路径查找支持不等于每个平台的渲染都已验证。

## 执行总览

Read the full skill on GitHub

Details

Skill name: auto-paper-reading
Repository: berglingtim/auto_paper_reading
License: MIT
Repo updated: 30 October 2026

FAQ

What does the Auto Paper Reading skill do?

How do I install the Auto Paper Reading skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add berglingtim/auto_paper_reading --skill auto-paper-reading. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Auto Paper Reading popular?

It lives in berglingtim/auto_paper_reading, which has 5 GitHub stars and 0 forks.

Related skills

Awesome Slam

Eyokurt/awesome-slam-skill

Architectural decision guide, algorithmic taxonomy, mathematical formulation, and engineering workflows for SLAM (Simultaneous Localization and Mapping), Visual Odometry (VO), Visual-Inertial Odometry (VIO), and LiDAR SLAM derived from the curated Awesome-SLAM collection and academic literature. Use when designing, selecting, implementing, or debugging SLAM systems, choosing between Monocular, Stereo, RGB-D, Visual-Inertial (ORB-SLAM3, VINS-Fusion, Basalt, SVO, DSO), or LiDAR pipelines (Cartographer, LOAM, LeGO-LOAM, LIO-SAM, FAST-LIO), configuring factor graph optimization (g2o, GTSAM, Ceres)

Research & Science

Six Layer Product Research

smith-horn/six-layer-product-research

Execute Teresa Torres' 6-Layer Problem Definition Framework for rigorous product research — from customer voice to scoped problem statement

Research & Science

Export Pharma Packaging Compliance

dungnotnull/export-pharma-packaging-compliance-agent-skill

Export Pharmaceutical Packaging Compliance Advisor - Pharmaceutical Packaging Regulatory Compliance evidence-backed analysis harness. Use this skill when analyzing pharmaceutical packaging compliance, regulatory requirements, export packaging standards, anti-moisture packaging, child-resistant/tamper-evident packaging, serialization/traceability (DSCSA, FMD), GMP requirements (ISO 15378), stability testing (ICH Q1A), or any pharmaceutical packaging regulatory compliance assessment. This skill provides evidenced, risk-disclosed analysis with domain-expert methodology and academic research integ

Research & Science

Create Research Experiment Repo

Shiba-hua/create-research-experiment-repo

为委托实验执行者开展工作的科研负责人，对齐报告、图表、目录和审批职责，获批后构建可追溯的实验管理仓库。适用于 Coding Agent、人类实验员或实验设备团队；不代替具体学科的实验设计与安全规程。

Research & Science

Med Mentor

Premansh12/med-mentor

Teach medical personnel (physicians, residents, medical students, nurses, PAs, paramedics) clinical knowledge and skills within this workspace, over multiple sessions, via short interactive HTML lessons. Use this whenever the user wants to study for board exams (USMLE, COMLEX, NCLEX, specialty boards), prepare for CME, build clinical reasoning skills (differentials, case work-ups), drill pharmacology/dosing, rehearse OSCE or procedural skills, or review any medical topic in a structured, spaced way — even if they don't say the word "teach" explicitly (e.g. "quiz me on antiarrhythmics", "help m

Research & Science

Clinical Eda Report

JeonKH81/MediStat-EDA

의학연구 tabular 데이터(.xlsx/.csv)에 대해 한국어 단일 HTML EDA 리포트를 자동 생성하는 스킬. 행이 관찰 단위(환자·내원·병변·검체 등), 열이 변수인 모든 의학연구 데이터셋이 대상이며 연구 디자인(후향/전향 코호트, RCT·임상시험, case-control, cross-sectional, registry, survey 등)을 가리지 않는다. n·변수 타입별 요약, 결측 패턴, 분포 플롯, 이상치(implausible value) 감지, 선택적 소그룹별 Table 1, 상관관계 heatmap, VIF를 모두 한 파일에 임베딩한다. 사용자가 임상연구·관찰연구·임상시험·환자 데이터·registry·연구 데이터셋·엑셀/CSV 파일을 업로드하면서 "EDA", "데이터 탐색", "탐색적 분석", "기초통계", "Table 1", "결측 보고", "분포 확인", "데이터 살펴봐", "데이터 점검" 같은 표현을 사용하면 적극적으로 트리거하라. 단순 통계 분석(t-test, Cox regression 등)이나 가설 검정 요청, 시각화 1개만 요청한 경우는 대상이 아니다. raw 영상(DICOM/JPEG)·ECG waveform·자연어 free tex

Research & Science

Browse the most popular Claude skills, all research & science skills, or write your own with the skills generator.