Skip to content
Back to search

student-llm-wiki

MarketplaceWorks with Claude CodeGitHub

📚 Student LLM Wiki — AI-compiled knowledge base for university students. Drop course slides, get a persistent interlinked wiki. Feynman review, exam prep, confidence decay, cross-course connections. Based on Karpathy's LLM Wiki pattern. Works with Claude Cowork / Claude Code / Claudian + Obsidian.学生大语言模型维基百科——专为大学生打造的AI编译知识库。

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/issacw228--student-llm-wiki/badge.svg)](/listings/issacw228--student-llm-wiki)

HTML

<a href="/listings/issacw228--student-llm-wiki"><img src="/api/listings/issacw228--student-llm-wiki/badge.svg" alt="findsafeskills safety rating" /></a>

Install

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 13 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "student-llm-wiki",
  3. "owner": {
  4. "name": "Ruizhe"
  5. },
  6. "plugins": [
  7. {
  8. "name": "student-llm-wiki",
  9. "source": "./",
  10. "description": "Student knowledge wiki — Karpathy LLM Wiki pattern with Feynman review, exam prep, confidence decay, and cross-course connections."
  11. }
  12. ]
  13. }
README.md · README · 382 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <div align="center">
  2. # 📚 Student LLM Wiki
  3. **用 AI 把课件变成你自己的知识库**
  4. **Turn your course slides into a personal wiki with AI**
  5. [中文](#中文) · [English](#english)
  6. </div>
  7. ---
  8. ## 中文
  9. ### 这是什么?
  10. 你把课件 PDF 丢进 `raw/` 文件夹,AI 自动帮你整理成结构化的 wiki 笔记,存在 `wiki/` 里。你用 [Obsidian](https://obsidian.md) 浏览这些笔记,用 AI 工具来复习、出题、检查薄弱环节。**你不需要手写任何笔记。**
  11. 基于 [Andrej Karpathy 的 LLM Wiki 模式](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f)。
  12. ---
  13. ### 你需要准备的东西
  14. **1. Obsidian(用来浏览笔记)**
  15. - 下载地址:[https://obsidian.md/download](https://obsidian.md/download)
  16. - 免费,支持 Windows / macOS / Linux / iOS / Android
  17. - 安装后还需要在社区插件里安装 **Dataview**(用于首页动态面板)
  18. **2. 选一个 AI 工具(任选其一即可)**
  19. | 工具 | 适合人群 | 获取方式 |
  20. |---|---|---|
  21. | Claude Code CLI | 熟悉命令行 | [安装说明](https://docs.anthropic.com/claude-code) |
  22. | Claude Code Web | 不想装任何东西 | [claude.ai/code](https://claude.ai/code) |
  23. | Cursor | 喜欢 IDE 界面 | [cursor.com](https://www.cursor.com) |
  24. | Trae | 国内用户 | [trae.ai](https://www.trae.ai) |
  25. | Cowork | 团队协作 | 参考 Cowork 文档 |
  26. | GitHub Copilot | 已有订阅 | VS Code 插件市场 |
  27. **3. 你的课件**(PDF 或其他文本格式)
  28. ---
  29. ### 第一步:下载项目,用 Obsidian 打开
  30. ```bash
  31. git clone https://github.com/IssacW228/student-llm-wiki.git
  32. cd student-llm-wiki
  33. ```
  34. 然后打开 Obsidian → 「打开本地仓库」→ 选择 `student-llm-wiki` 文件夹。
  35. > **必须安装 Dataview 插件**:Obsidian 设置 → 社区插件 → 浏览 → 搜索 `Dataview` → 安装并启用。
  36. ---
  37. ### 第二步:选择你的 AI 工具并打开项目
  38. <details>
  39. <summary><strong>Claude Code CLI(命令行)</strong></summary>
  40. 1. 确保已安装 Claude Code CLI 并登录
  41. 2. 在项目目录运行:
  42. ```bash
  43. claude
  44. ```
  45. 3. 配置文件(`.claude/`)会自动加载,无需额外设置
  46. 4. 直接输入命令或自然语言即可开始
  47. </details>
  48. <details>
  49. <summary><strong>Claude Code Web(网页版,无需安装)</strong></summary>
  50. 1. 打开 [claude.ai/code](https://claude.ai/code)
  51. 2. 将此仓库连接到 Claude Code
  52. 3. 配置文件(`.claude/`)自动识别,直接使用
  53. </details>
  54. <details>
  55. <summary><strong>Cursor</strong></summary>
  56. 1. 用 Cursor 打开 `student-llm-wiki` 文件夹
  57. 2. 规则文件已在 `.cursor/rules/wiki.mdc`,自动生效
  58. 3. 在 Cursor 聊天框直接输入命令,无需额外配置
  59. </details>
  60. <details>
  61. <summary><strong>Trae</strong></summary>
  62. 1. 用 Trae 打开 `student-llm-wiki` 文件夹
  63. 2. Skills 已在 `.trae/skills/`,自动识别
  64. 3. 在对话框直接输入命令
  65. </details>
  66. <details>
  67. <summary><strong>Cowork</strong></summary>
  68. 1. 将 Cowork 项目指向 `student-llm-wiki` 文件夹
  69. 2. `COWORK-INSTRUCTIONS.md` 会自动加载
  70. 3. 直接开始对话
  71. </details>
  72. <details>
  73. <summary><strong>GitHub Copilot / OpenAI Codex</strong></summary>
  74. 1. 用支持 Copilot 的编辑器打开项目文件夹
  75. 2. `AGENTS.md` 会被自动读取
  76. 3. 直接在对话框输入命令
  77. </details>
  78. ---
  79. ### 第三步:导入你的第一份课件
  80. 1. 在 `raw/` 下新建文件夹,用你的课程代号命名(如 `raw/MATH1001/`)
  81. 2. 把课件 PDF 复制进去
  82. 3. 在 AI 工具里输入:
  83. ```
  84. ingest raw/MATH1001/L1.pdf
  85. ```
  86. 4. AI 会自动生成概念页、课程总览、来源摘要
  87. 5. 回到 Obsidian,打开 `Home.md` 查看结果
  88. > 同一份文件不会被重复导入(通过文件 hash 去重)。
  89. ---
  90. ### 命令速查
  91. | 命令 | 作用 |
  92. |---|---|
  93. | `ingest raw/XXXX/L1.pdf` | 导入课件,自动生成笔记、课程总览、跨课连接记录 |
  94. | `lint` | 检查笔记健康度:断链、孤立页、过期概念、缺失总览、术语表补全 |
  95. | `review XXXX` | 费曼复习模式:AI 提问,你来回答 |
  96. | `exam-prep XXXX` | 根据薄弱概念自动出练习题 |
  97. | `diagram ConceptName` | 为某个概念页添加 Mermaid 图解(流程图/架构图/时序图等) |
  98. 也可以用自然语言,比如「帮我复习这门课」或「把这个文件导入知识库」。
  99. ---
  100. ### 添加新课程
  101. 1. 在 `raw/` 下创建对应课程文件夹(如 `raw/PHYS1001/`)
  102. 2. 把课件放进去
  103. 3. 运行 `ingest`,课程总览页会自动创建
  104. 4. `Home.md` 首页会动态显示新课程,无需手动修改
  105. ---
  106. ### 项目结构
  107. ```
  108. student-llm-wiki/
  109. ├── raw/ ← 放课件(只读,AI 不会修改这里)
  110. │ └── XXXX/ ← 按课程分文件夹
  111. ├── wiki/ ← AI 生成的笔记(自动维护)
  112. │ ├── concepts/ ← 概念页(每概念一页,含图解)
  113. │ ├── courses/ ← 课程总览(ingest 时自动创建)
  114. │ ├── sources/ ← 课件摘要
  115. │ ├── exam-prep/ ← 练习题
  116. │ ├── connections-log.md ← 跨课程连接发现记录
  117. │ ├── contradictions.md ← 矛盾与张力追踪
  118. │ ├── glossary.md ← 中英术语表(自动填充)
  119. │ ├── overview.md ← 跨课程综述(自动更新)
  120. │ ├── index.md ← 总目录
  121. │ └── hot.md ← AI 上下文缓存(每次 session 首读)
  122. ├── Home.md ← Obsidian 首页仪表盘
  123. │
  124. │ 以下为 AI 工具配置文件,无需修改:
  125. ├── .claude/ ← Claude Code 配置
  126. ├── .cursor/ ← Cursor 配置
  127. ├── .trae/ ← Trae 配置
  128. ├── AGENTS.md ← Copilot/Codex 配置
  129. └── COWORK-INSTRUCTIONS.md ← Cowork 配置
  130. ```
  131. ---
  132. ### 致谢
  133. - [Andrej Karpathy's LLM Wiki](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f) — 原始模式
  134. - [claude-obsidian](https://github.com/AgriciDaniel/claude-obsidian) — hot cache、manifest 去重、skill 拆分
  135. - [Obsidian](https://obsidian.md) — 知识平台
  136. ### 许可证
  137. MIT
  138. ---
  139. ## English
  140. ### What is this?
  141. You drop course PDF slides into the `raw/` folder. The AI automatically organizes them into structured wiki notes in `wiki/`. You browse the notes with [Obsidian](https://obsidian.md), and use AI tools to review, generate practice questions, and surface weak spots. **You never write notes yourself.**
  142. Based on [Andrej Karpathy's LLM Wiki pattern](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f).
  143. ---
  144. ### Prerequisites
  145. **1. Obsidian (for browsing your notes)**
  146. - Download: [https://obsidian.md/download](https://obsidian.md/download)
  147. - Free. Available on Windows / macOS / Linux / iOS / Android
  148. - After installing, also install the **Dataview** community plugin (required for the Home dashboard)
  149. **2. An AI tool (pick just one)**
  150. | Tool | Best for | Get it |
  151. |---|---|---|
  152. | Claude Code CLI | Terminal users | [Install guide](https://docs.anthropic.com/claude-code) |
  153. | Claude Code Web | No install needed | [claude.ai/code](https://claude.ai/code) |
  154. | Cursor | IDE-style interface | [cursor.com](https://www.cursor.com) |
  155. | Trae | — | [trae.ai](https://www.trae.ai) |
  156. | Cowork | Team use | See Cowork docs |
  157. | GitHub Copilot | Existing subscribers | VS Code marketplace |
  158. **3. Your course slides** (PDF or other text formats)
  159. ---
  160. ### Step 1: Clone the repo and open in Obsidian
  161. ```bash
  162. git clone https://github.com/IssacW228/student-llm-wiki.git
  163. cd student-llm-wiki
  164. ```
  165. Open Obsidian → "Open folder as vault" → select the `student-llm-wiki` folder.
  166. > **Install the Dataview plugin**: Obsidian Settings → Community plugins → Browse → search `Dataview` → Install and enable.
  167. ---
  168. ### Step 2: Open the project with your AI tool
  169. <details>
  170. <summary><strong>Claude Code CLI</strong></summary>
  171. 1. Make sure Claude Code CLI is installed and you're logged in
  172. 2. Run in the project folder:
  173. ```bash
  174. claude
  175. ```
  176. 3. The config folder (`.claude/`) is auto-detected — no extra setup needed
  177. 4. Type commands or plain English to get started
  178. </details>
  179. <details>
  180. <summary><strong>Claude Code Web (no install needed)</strong></summary>
  181. 1. Go to [claude.ai/code](https://claude.ai/code)
  182. 2. Connect this repository to Claude Code
  183. 3. The `.claude/` config is auto-detected — start typing commands directly
  184. </details>
  185. <details>
  186. <summary><strong>Cursor</strong></summary>
  187. 1. Open the `student-llm-wiki` folder in Cursor
  188. 2. Rules are pre-configured in `.cursor/rules/wiki.mdc` — auto-applied
  189. 3. Type commands in the Cursor chat panel
  190. </details>
  191. <details>
  192. <summary><strong>Trae</strong></summary>
  193. 1. Open the `student-llm-wiki` folder in Trae
  194. 2. Skills are pre-configured in `.trae/skills/` — auto-detected
  195. 3. Type commands in the chat panel
  196. </details>
  197. <details>
  198. <summary><strong>Cowork</strong></summary>
  199. 1. Point your Cowork project at the `student-llm-wiki` folder
  200. 2. `COWORK-INSTRUCTIONS.md` loads automatically
  201. 3. Start chatting
  202. </details>
  203. <details>
  204. <summary><strong>GitHub Copilot / OpenAI Codex</strong></summary>
  205. 1. Open the project in an editor with Copilot support
  206. 2. `AGENTS.md` is read automatically
  207. 3. Type commands in the chat panel
  208. </details>
  209. ---
  210. ### Step 3: Ingest your first slide deck
  211. 1. Create a folder under `raw/` named after your course (e.g. `raw/MATH1001/`)
  212. 2. Copy your PDF into that folder
  213. 3. In your AI tool, type:
  214. ```
  215. ingest raw/MATH1001/L1.pdf
  216. ```
  217. 4. The AI generates concept pages, a course overview, and a source summary
  218. 5. Switch to Obsidian and open `Home.md` to see your new notes
  219. > The same file won't be ingested twice — files are tracked by hash.
  220. ---
  221. ### Commands
  222. | Command | What it does |
  223. |---|---|
  224. | `ingest raw/XXXX/L1.pdf` | Import a slide deck — generates notes, course overview, and cross-course link records |
  225. | `lint` | Check wiki health: broken links, orphan pages, stale concepts, missing overviews, glossary gaps |
  226. | `review XXXX` | Feynman review mode: AI quizzes you |
  227. | `exam-prep XXXX` | Auto-generate practice questions from weak concepts |
  228. | `diagram ConceptName` | Add a Mermaid diagram to a concept page (flowchart / architecture / sequence / etc.) |
  229. You can also use plain English: "quiz me on this course" or "import this file into the wiki".
  230. ---
  231. ### Adding a new course
  232. 1. Create a folder under `raw/` for the course (e.g. `raw/PHYS1001/`)
  233. 2. Put your slides in it
  234. 3. Run `ingest` — a course overview page is created automatically
  235. 4. The `Home.md` dashboard updates dynamically — no manual edits needed
  236. ---
  237. ### Project structure
  238. ```
  239. student-llm-wiki/
  240. ├── raw/ ← Your slides (read-only — AI never modifies this)
  241. │ └── XXXX/ ← One folder per course
  242. ├── wiki/ ← AI-generated notes (auto-maintained)
  243. │ ├── concepts/ ← Concept pages (one per concept, with diagrams)
  244. │ ├── courses/ ← Course overviews (auto-created on ingest)
  245. │ ├── sources/ ← Slide summaries
  246. │ ├── exam-prep/ ← Practice questions
  247. │ ├── connections-log.md ← Cross-course link discovery log
  248. │ ├── contradictions.md ← Contradiction & tension tracker
  249. │ ├── glossary.md ← Bilingual term index (auto-populated)
  250. │ ├── overview.md ← Cross-course synthesis (auto-updated)
  251. │ ├── index.md ← Master catalog
  252. │ └── hot.md ← AI context cache (read first each session)
  253. ├── Home.md ← Obsidian dashboard
  254. │
  255. │ AI tool config — no need to touch these:
  256. ├── .claude/ ← Claude Code config
  257. ├── .cursor/ ← Cursor config
  258. ├── .trae/ ← Trae config
  259. ├── AGENTS.md ← Copilot/Codex config
  260. └── COWORK-INSTRUCTIONS.md ← Cowork config
  261. ```
  262. ---
  263. ### Credits
  264. - [Andrej Karpathy's LLM Wiki](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f) — the original pattern
  265. - [claude-obsidian](https://github.com/AgriciDaniel/claude-obsidian) — hot cache, manifest dedup, skill decomposition
  266. - [Obsidian](https://obsidian.md) — the knowledge platform
  267. ### License
  268. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.