메뉴
从论文 PDF Workflow：本地 MinerU 提取 Markdown，在 Cursor/Codex 会话内生成 Marp 幻灯（不使用额外 LLM/生图 API），再本地导出 pptx/PDF。Use when the user wants paper-to-slides, paper to PPT, thesis deck, or Marp export without OpenAI/Gemini/OpenRouter keys.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Create, edit, replicate, and iteratively refine editable research and technical diagrams in diagrams.net/draw.io (.drawio XML) from prompts, papers, repositories, screenshots, or existing diagrams. Use when asked to generate a scientific figure, paper method diagram, ML/system architecture diagram, draw.io file, reference figure reproduction, browser screenshot feedback loop, Windows-safe draw.io preview, or layout fixes for text overlap, arrows, colors, icons, fonts, and component alignment.
Always use when user asks to create, generate, draw, or design a diagram, flowchart, architecture diagram, ER diagram, sequence diagram, class diagram, network diagram, mockup, wireframe, or UI sketch, or mentions draw.io, drawio, drawoi, .drawio files, or diagram export to PNG/SVG/PDF.
Use when the user requests diagrams, flowcharts, architecture diagrams, ER diagrams, UML / sequence / class diagrams, SysML / MBSE diagrams (block definition, internal block, requirement, parametric), BPMN business process diagrams, swimlane / cross-functional flowcharts, network topology, cloud architecture from Terraform or Kubernetes manifests, ML/DL model figures (Transformer/CNN/LSTM), mind maps, or any visualization. Also use proactively when explaining systems with 3+ components, complex data flows, or relationships that benefit from visual representation. Best suited when the diagram needs custom styling, rich shape vocabulary, swimlanes, or exportable images (PNG/SVG/PDF/JPG). Generates .drawio XML and exports locally via the native draw.io desktop CLI.
Orchestrates AI research agent workflows (paper deep-dive, survey builder, cite verify, repro pack) with durable delivery to Feishu/Lark, Notion, or Obsidian/Markdown. Use when the user asks for literature workflows, paper cards, survey/taxonomy, citation verification, reproducibility packages, cross-platform research documentation, or mentions AI Research Skills / 科研 skill 工作流.
Write source-grounded research experiment reports and experiment-log entries for ML/vision/graphics/robotics projects. Default durable destination is Notion; Feishu/Lark or Obsidian/Markdown only when the user explicitly names them. Use when the user asks to整理/撰写/补全/审查实验报告, 实验记录, 训练报告, 评测报告, ablation report, baseline report, reproduction report, Notion experiment pages, or research-document experiment summaries.
Notion document and database workflow for reading, editing, structuring, and verifying pages used by research skills (paper cards, deep dives, surveys, meeting notes, bilingual archives). Use when the user provides a notion.so URL, names Notion as the destination, asks to create or update a Notion page/database, or research-doc-workflow routes a durable write to Notion. Applies Notion-native format deltas so Feishu/Obsidian syntax does not leak into reader-facing content.
| name | paper-to-slides |
| description | 从论文 PDF Workflow：本地 MinerU 提取 Markdown，在 Cursor/Codex 会话内生成 Marp 幻灯（不使用额外 LLM/生图 API），再本地导出 pptx/PDF。Use when the user wants paper-to-slides, paper to PPT, thesis deck, or Marp export without OpenAI/Gemini/OpenRouter keys. |
Short description: 论文 PDF → Marp（零额外交互 API）。Do not invoke or configure any additional external model API for this workflow.
.tools/mineru-md.sh）；脚本保存在 vault 内，Python 环境默认保存在本机 ~/Library/Application Support/WorldModelVault/envs/，与 Vault Setup.md 一致。不要使用 HKUDS/Paper2Slides 等默认依赖云端 Key 的一键工具，除非用户明确说已自行改为全本地后端（本 skill 不展开）。
用户在做：组会幻灯、读书报告、论文分享、把项目中的 PDF 变成可讲 deck，且希望 skill 化、可重复、不绑新 API。
3-Projects/<项目>/assets/mineru-out/<论文短名>/ 或 2-Learnings/EmbodiedWorld/assets/mineru-out/<短名>/3-Projects/<项目>/decks/YYYY-MM-DD-<论文短名>.md（或同级 decks/ 下）.md 同目录的 exports/，避免和 Obsidian 笔记 clutter若用户已在 3-Projects/.../Paper Cards.md 有 paper card，可优先对齐其中的标题、链接与配图路径。
确认输入
PDF 路径；听众（导师/同门/外行）；时长或页数上限；语言（中文/英文）。
本地提取（若尚无可靠 Markdown）
在 vault 根执行：
.tools/mineru-md.sh "/path/to/paper.pdf" "3-Projects/<项目>/assets/mineru-out/<短名>"
在输出目录中找到主 Markdown（以 MinerU 实际结构为准，常为 *.md 或子目录中的 md）。
在会话内理解论文
阅读 MinerU md（与必要图片路径）；整理：问题、方法、实验、结论、限制；标出必须在幻灯上出现的图（用相对 Marp 文件的路径或复制/引用到 decks/assets/）。
生成 Marp 文稿
新建或更新 decks/...md，包含：
marp: true、theme（若用户无偏好用 default 或 gaia）、paginate: true、size: 16:9--- 分页；标题页、大纲、贡献/方法、实验、结论、备份页（可选），确保相对 Marp 文件可解析导出（用户本机，Agent 只给命令）
marp decks/YYYY-MM-DD-foo.md --pptx -o exports/YYYY-MM-DD-foo.pptx
marp decks/YYYY-MM-DD-foo.md --pdf -o exports/YYYY-MM-DD-foo.pdf
质量检查清单（口头提醒用户）
字体是否嵌入、公式是否需截图、图是否模糊、是否需加一句「复现/数据来自原文表 X」。
---
marp: true
theme: default
paginate: true
size: 16:9
---
# 论文标题
作者（会议 / arXiv） · 你的名字 · YYYY-MM-DD
---
# 我们在讲什么
- 问题：…
- 挑战：…
---
# 方法概览
- …
---
# 实验与结论
- …
#### 标题块）：适合文献档案与精读笔记。assets/ 图，但不要互相覆盖「唯一真源」：MinerU md 保持提取层；Marp 保持演讲层。| 需求 | 做法 |
|---|---|
| 只要提纲、不要文件 | 仍按步骤 3 输出分页大纲，由用户贴进 Keynote |
| PDF 已是幻灯式逐页 | 考虑矢量保留的 pdf→pptx 工具；本 skill 仍以「论文全文」为主场景 |
| 公式很多 | Marp 有限；建议关键式一行展示，其余放附录页或截图 |