juliye2025/evil-read-arxiv

extract-paper-images

从 arXiv ID/link、公开直接 PDF URL 或本地论文 PDF 提取架构图、方法图和实验图，保存到 Obsidian images 目录并生成索引。Use when the user asks to extract figures or another research skill needs embeddable images.

Source text: Chinese

paper-analyze

深度分析 arXiv 论文、公开 PDF URL、本地 PDF、研究项目页、技术博客或普通网页，自动识别来源类型并生成带证据与图片的 Obsidian 笔记。Use when the user provides an arXiv ID/link、public PDF URL、company research page、blog URL、local PDF or existing note and asks for analysis, summary, evaluation, or knowledge-base…

Source text: Chinese

conf-papers

通过 DBLP 与 Semantic Scholar 搜索 CVPR、ICCV、ECCV、ICLR、AAAI、NeurIPS、ICML 等顶会论文，评分后生成 Obsidian 年度推荐笔记。Use when the user asks for a conference/year paper list, top-conference recommendations, or papers filtered by venue.

Source text: Chinese

paper-search

在本地 Obsidian Vault 的论文笔记中按标题、作者、关键词、标签、领域或 arXiv ID 搜索并排序结果。Use when the user asks to find, filter, or revisit papers already captured in local notes; do not use for downloading new papers.

Source text: Chinese

start-my-day

生成每日 arXiv 与 Semantic Scholar 论文推荐并写入 Obsidian Vault。Use when the user asks to start the day, create today's paper brief, or generate recommendations for a specified date.

Source text: Mixed languages