20:"$Sreact.view_transition" 21:I[9754,["/_next/static/chunks/2t36f5p4pjzlb.js","/_next/static/chunks/1gyqeidjm5eoa.js","/_next/static/chunks/0_fjnxc4tt_l_.js","/_next/static/chunks/1ymt1shyhmu-4.js","/_next/static/chunks/3kv2zf0i7s2wc.js","/_next/static/chunks/1vs_3g_g75zmk.js","/_next/static/chunks/2g4dh802s8oz3.js","/_next/static/chunks/1bc3f0shukf1j.js","/_next/static/chunks/192kgs9v_t14j.js","/_next/static/chunks/0z6pw0v5c_lx5.js","/_next/static/chunks/08r60e6icmtkc.js","/_next/static/chunks/3trx6c0i7_x-3.js","/_next/static/chunks/40sro5bzp9bi_.js","/_next/static/chunks/1wgf3tk38y6cf.js"],"ProjectDetail"] 22:T3cc4, understand-anything — 把任何代码变成可探索的知识图谱

egonex-ai · understand-anything

图形的目的在于教会人，而不是让人惊叹。

Graphs that teach > graphs that impress.

INSPECT

扫描一次，之后就能一直追问。

SCAN ONCE · ASK FOREVER

一个多智能体流水线把整个项目拆开，再拼回一张以文件、函数、类与依赖为节点的知识图谱。你不再需要通读每一行，而是直接走进这张结构。

仪表盘提供引导式导览与模糊 + 语义搜索。找不到精确词也能带着意图来问，图谱替你找到它们之间的连接。

还能评估当前改动的影响——在提交之前先看见这一处会牵动哪些依赖与调用点。

角色自适应的界面与分层可视化，让新手看概貌、老手看纵深，同一张图谱有不同的读法。

PLATFORMS

住在你已经在用的工具里。

读懂项目的知识不锁在一处。作为插件，它进驻多种 AI 编码环境，让你在原处展开知识图谱——Claude Code、Codex、Cursor、Copilot、Gemini CLI 等。

Claude Code

在你熟悉的会话旁长出可探索的结构图。

Codex

把散落的代码收敛成内聚星簇。

Cursor / Copilot

在编辑器内直接提问与检索。

Gemini CLI · opencode

同一分析流水线的更多入口。

MIT · TypeScript · 单源分析 · 数据目录已收敛为 .ua/ 并保持向后兼容

RECURSION

不只代码——让它教会你知识本身。

分析的对象并不局限于代码。它也能读取Karpathy 模式的知识库（LLM wiki），把它组织成一张带社区聚类的力导向图谱——让一群高度相关的话题自然聚成肉眼可辨的群。

最新的设计图模式更可以把 Figma 文件映射成可导航的知识图谱。看懂代码、看懂文档、也看懂一块设计的构成。

PROGRESSION

从一句「依赖在哪」到读完整个项目。

→

新成员的第一天——生成一份引导式上手指南，从团队的代码里学会团队的约定。

→

改动前的最后一问——分析当前改动的影响范围，再决定要怎么改。

→

持续学习——只分析变动的文件（增量默认），提交后自动更新，随时可重跑。

→

巨型 monorepo 也能入口——把范围收敛到某个子目录，海量代码也不必一次吞下。

understand-anything · egonex-ai · MIT · TypeScript understand-anything.com GitHub

23:T3ca3, understand-anything — Turn any code into an explorable knowledge graph

egonex-ai · understand-anything

Graphs that teach, more than graphs that impress.

Turn any code into an explorable knowledge graph.

INSPECT

Scan once, then keep asking.

SCAN ONCE · ASK FOREVER

A multi-agent pipeline takes a project apart, then rebuilds it as a knowledge graph of files, functions, classes, and dependencies. You no longer read every line — you step inside the structure.

The board pairs guided tours with fuzzy + semantic search. You can come with an intent rather than an exact phrase, and the graph shows you the connections beneath it.

You can also weigh the impact of your current changes — seeing which dependencies and call-sites a single edit would tug at, before you commit.

Persona-adaptive views and layered visualization give newcomers a map and veterans the depth — the same graph, read several ways.

PLATFORMS

Live inside the tools you already use.

This knowledge isn't locked in one place. As a plugin it enters several AI coding environments, unfolding the graph where you already work — Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.

Claude Code

An explorable structure grows beside your conversations.

Codex

Scattered code folds toward a coherent cluster.

Cursor / Copilot

Search and ask from inside the editor.

Gemini CLI · opencode

More entrances to the same analysis pipeline.

MIT · TypeScript · single-source analysis · the data directory is now .ua/, fully backward compatible

RECURSION

Not only code — let it teach you the knowledge.

The subject isn't limited to code. It can also read a Karpathy-pattern knowledge base (LLM wiki) and organize it into a force-directed graph with community clustering — letting clusters of related topics surface as constellations you can see at a glance.

A newer design mode can map Figma files into a navigable knowledge graph as well. Understand code, understand documentation, and see how a design is put together.

PROGRESSION

From "where does this depend on?" to reading the whole project.

→

Day one for a newcomer — generate a guided onboarding guide and learn your team's conventions from the team's own code.

→

The last question before a change — analyze the blast radius of your current edits before deciding how to make them.

→

Keep learning — it re-analyzes only changed files (incremental by default), auto-updates after each commit, and can be re-run anytime.

→

Huge monorepos still find a way in — scope analysis to a subdirectory so massive code doesn't have to be swallowed whole.

understand-anything · egonex-ai · MIT · TypeScript understand-anything.com GitHub

a:[["$","$20",null,{"enter":{"secondary-forward":"secondary-forward","secondary-back":"secondary-back-enter","default":"none"},"exit":{"secondary-forward":"secondary-forward","secondary-back":"secondary-back-exit","default":"none"},"update":{"secondary-forward":"secondary-forward","secondary-back":"secondary-back-update","default":"none"},"default":"none","children":["$","$L21","cmsn448ag000p1s16rirf9agn:zh-CN:beautified",{"project":{"id":"cmsn448ag000p1s16rirf9agn","slug":"egonex-ai-understand-anything","name":"understand-anything","description":"将代码库、知识库或文档转化为交互式知识图谱","authorName":"egonex-ai","authorVerification":{"verified":false,"verifiedAt":null},"category":"development_tools","subcategory":"data_ai_tools","experienceUrl":"https://understand-anything.com/","qualityScore":100,"featured":false,"officialSite":null,"views":{"original":{"available":true},"official":{"available":false,"reason":"disabled_by_policy"},"beautified":{"available":true,"reason":null},"defaultView":"beautified"},"beautifiedView":{"revisionId":"cmt2zgzs500gu627163t6p1ty","kind":"artifact","artifact":{"schemaVersion":"localized-static-html-v4","kind":"static_html","locales":["zh-CN","en"],"defaultLocale":"zh-CN","documents":{"zh-CN":"$22","en":"$23"},"designDirections":{"schemaVersion":"dodot-design-directions-v1","directions":[{"id":"orbit-reader","name":"Orbit Reader","projectMetaphor":"分析管道是一条把原始代码库牵引进星图的传送带——一次解析、不断重连、持续扩张的深度阅读轨道","signatureVisual":"一条从左侧原始文件簇滑向右上方星群的单行连接轨道，轨道上按顺序排布三个发光节点（文件→函数→依赖），节点间距体现知识的收缩与凝集","heroGeometry":"一条水平倾角约8度、贯穿整个首屏又略带弧度的连接轨道，左侧为散点群、右侧收敛为内聚的多边形星簇，轨道两侧仅有少量带标签的箭头标注解析阶段","sectionSequence":["slogan-pipeline","orbit-flow-stages","inspect-and-ask","platform-inlays","recursion-teaching","onboard-progression"],"surfaceModel":"深浅分层的代码簿页：轨道与节点浮于一张深色羊皮纸质感之上，文字以浅色呈现，节点用低饱和高对比的金绿色敲击点，无大块色板仅用细分割线","palette":"近黑墨纸底、暖白文字、金绿节点、暮紫连接线、单一强调青用于阶段编号","typography":"等宽字体贯穿标题与数据标注，具非常规字距；正文衬线轻古卷味，强调阅读感而非科技炫技","desktopComposition":"Hero 中轨道占左侧2/3，右侧1/3竖排三行解析结果摘要与一条可点击的“开始阅读”轨迹；阶段以连续的错落小标尺呈现而非卡片格","mobileTransformation":"轨道旋转为纵向阶梯落差，节点自上而下依次点亮，点击任一节点展开对应阶段的简要说明，版面收敛为单列流水","mediaStrategy":"无外部媒体；标志视觉用纯 CSS 的连接轨道与节点动画表达解析→凝集关系","forbiddenPatterns":["三张指标卡并排","浏览器/编辑器/终端外壳假界面","发光粒子网格充当背景","截断的假截图或远程图片"],"compositionFingerprint":"single ascending pipeline orbit with ordered stages, left scatter to right cluster, no card grid"},{"id":"knowledge-vault","name":"Knowledge Vault","projectMetaphor":"把整座代码库折叠成一间可进入的知识书房——书架、索引卡与提问台代替单调的控制台","signatureVisual":"一只发光的钥匙孔形翻阅台：上半是半开的书本展开成星图，下半是提问槽，一道光线从书页投向台面形成“探索”路径","heroGeometry":"两分构图：左侧立起的半开书脊，书页翻开的轮廓向外延展成连接的节点网络，右侧留白处仅有排布成切线方向的提问语句，形成阅读台的景深","sectionSequence":["vault-cover","bookshelf-index","ask-desk","search-and-diff","wiki-clusters","learning-colophon"],"surfaceModel":"图书室材质：木纹书桌、毛边纸书页、细铜镶边；区块层层叠叠如书架隔层，用抽屉式开合与纸张翻页过渡","palette":"暖栗木色、奶油纸白、铜金强调、墨绿书脊、沉红书签","typography":"高衬线书卷标题搭配低对比正文衬线，行高慷慨；章节编号仿图书章节用罗马数字","desktopComposition":"左书右台的雕塑式两栏，右侧提问句逐行点亮；书架索引以期刊抽卡列而非卡片网格，底部一排附录脚注呼应知识库","mobileTransformation":"书本立起来变为上下翻页的书页块，提问台下沉为收拢的对话槽，书架索引改为可滑动抽拉的横卷","mediaStrategy":"无外部媒体；以开卷与钥匙孔轮廓的 CSS 线稿表达提取→理解","forbiddenPatterns":["终端或编辑器窗口壳","均质功能卡片墙","发光的装饰网格背景","假制的表单控件或搜索框"],"compositionFingerprint":"sculptural book-and-desk diptych with roman-numeral shelf index and dialed question column, no metric grid"},{"id":"recon-map","name":"Recon Map","projectMetaphor":"代码库被当作一片待勘察的领地——先扫描全景、再逐层挖掘、最终在原地留下可追问的路标","signatureVisual":"一幅按纵深剖开的立体等高线地图：地表是概览层，往下逐层切开暴露出函数与依赖层的等高线，每层有发光的锚点","heroGeometry":"斜剖视等距构图：多层半透明的水平切面自底部向顶部叠放，切面间用虚线连接表示依赖穿越，右上角一枝罗盘针指示“入口”，俯瞰的等高线代替市场首屏","sectionSequence":["terrain-overview","stratigraphy-layer-cuts","landmark-nodes","impact-ripple","platform-territories","trail-guide"],"surfaceModel":"测绘剖面图：半透明切面、细网格辅助线仅存在于切面内部、等高线圈、罗盘刻度；区块如地质分层上下堆叠切开","palette":"深绿测绘图底、纸白等高线、琥珀锚点、锈红影响波纹、雾蓝背景层","typography":"粗体无衬线作为图层标签，小号等宽附注作为坐标读数；章节号用图层序号「L1」「L2」呈现","desktopComposition":"Hero 用斜剖视二层切面占据主景，右侧纵列坐标图例；图层章节以不同高度切面纵向堆叠而非左右卡片，影响章节以同心扩散波纹呈现","mobileTransformation":"斜剖视拆为逐层下滑的横剖条，每滑到一层切面展开为全宽剖面，锚点并列于侧缘引导点击","mediaStrategy":"无外部媒体；以等高线、罗盘与切面层动画表现勘察→挖掘→标图","forbiddenPatterns":["营销大标题配三指标","浏览器壳或仪表盘壳冒充","粒子/辉光背景填充","并列的均质功能卡"],"compositionFingerprint":"cutaway stratigraphic terrain with layered translucent slices, compass entry, concentric impact ripples, no flat card wall"}],"selectedDirectionId":"orbit-reader","selectionRationale":"奥比特阅读者方向最贴近项目“编码到交互式知识图谱，可搜索可提问”的机制：把散点原始代码收敛为内聚星簇这一构图直接物化了描述中“代码→知识图”的核心关系，且首屏即用解析阶段顺序传达多智能体分析流水线的事实，而非营销式指标。其逐节点凝集的运动、约束在单行轨道上的可解释步骤，能同时承载搜索与提问两类交互，且完全不依赖浏览器/编辑器外壳或伪造截图。相较知识库书架方向（更偏体验氛围）与测绘地图方向（更偏空间叠层），该方向在证据契合、机制表达和移动端阶梯化重构上最为清晰，风险也最低。"},"designBrief":{"schemaVersion":"dodot-design-brief-v1","selectedDirectionId":"orbit-reader","projectMetaphor":"分析管道是一条把原始代码库牵引进星图的传送带——一次解析、不断重连、持续扩张的深度阅读轨道","signatureVisual":"一条从左侧原始文件簇滑向右上方星群的单行连接轨道，轨道上按顺序排布三个发光节点（文件→函数→依赖），节点间距体现知识的收缩与凝集","compositionFingerprint":"single ascending pipeline orbit with ordered stages, left scatter to right cluster, no card grid","heroGeometry":"一条水平倾角约8度、贯穿整个首屏又略带弧度的连接轨道，左侧为散点群、右侧收敛为内聚的多边形星簇，轨道两侧仅有少量带标签的箭头标注解析阶段","sectionSequence":["slogan-pipeline","orbit-flow-stages","inspect-and-ask","platform-inlays","recursion-teaching","onboard-progression"],"surfaceModel":"深浅分层的代码簿页：轨道与节点浮于一张深色羊皮纸质感之上，文字以浅色呈现，节点用低饱和高对比的金绿色敲击点，无大块色板仅用细分割线","palette":"近黑墨纸底、暖白文字、金绿节点、暮紫连接线、单一强调青用于阶段编号","typography":"等宽字体贯穿标题与数据标注，具非常规字距；正文衬线轻古卷味，强调阅读感而非科技炫技","desktopComposition":"Hero 中轨道占左侧2/3，右侧1/3竖排三行解析结果摘要与一条可点击的“开始阅读”轨迹；阶段以连续的错落小标尺呈现而非卡片格","mobileTransformation":"轨道旋转为纵向阶梯落差，节点自上而下依次点亮，点击任一节点展开对应阶段的简要说明，版面收敛为单列流水","mediaStrategy":"无外部媒体；标志视觉用纯 CSS 的连接轨道与节点动画表达解析→凝集关系","evidencePlan":["以描述开头‘Graphs that teach > graphs that impress’建立 slogan 与价值主张，首屏即时传达","多智能体分析流水线由 README 的 Features 与‘Analyze your codebase / Explore the dashboard’列出，作为阶段轨道的事实依据","‘可搜索、可提问’与‘分析当前变更的影响’由概览与 README 的 Ask anything 一节支撑，作为 inspect-and-ask 与 impact 内容来源","支持 Claude Code、Codex、Cursor、Copilot、Gemini CLI 等平台由描述与概览支撑，作为 platform-inlays 内容来源","Karpathy 模式的 LLM wiki 知识库与力导向社区聚类由证据 claim 支撑，作为 recursion-teaching 内容来源"],"localeContract":{"zh-CN":"把任何代码变成可探索的知识图谱","en":"Turn any code into an explorable knowledge graph"},"forbiddenPatterns":["三张指标卡并排作为首屏主体","浏览器/编辑器/终端/仪表盘外壳冒充产品界面","发光粒子网格或装饰渐变替代项目特有视觉","伪造搜索框、开关、假截图或假测试数据","仅换色/圆角/模糊的方案且未变隐喻、标志视觉、Hero 几何与移动重构"]},"visualReview":{"schemaVersion":"dodot-automated-design-review-v1","status":"approved","summary":"页面忠实实现了 orbit-reader 项目专属 Brief：首屏即落 slogan 与单行连接轨道标志视觉(左侧散点文件簇->三个发光节点文件/函数/依赖->右上方收敛星群)，六段章节顺序与 sectionSequence 完全对应；所有事实可追溯至 material.json 的描述/README/发布说明 claims，无据造物；无三卡指标墙、无设备壳、无装饰粒子网格替代标志视觉、无假控件；双语章节等价且自然。仅一处非阻断提示：Brief 的移动端规格要求轨道旋转为纵向阶梯并支持点击节点展开，页面在移动端仅将版式重构为单列与纵向阶段条，标志轨道本身未旋转、节点未支持点击展开。","checks":{"projectFit":{"passed":true,"evidence":"sectionSequence [slogan-pipeline,orbit-flow-stages,inspect-and-ask,platform-inlays,recursion-teaching,onboard-progression] 全部实现，data-dodot-section 分别对应 INSPECT/PLATFORMS/RECURSION/PROGRESSION；Hero 首屏即 slogan(Graphs that teach>...)+右侧竖排三行解析摘要+可点击开始阅读轨迹，符合 desktopComposition；副标题把 any code 折叠成可搜索可提问的图谱一以贯之，无偏离隐喻。"},"evidenceBound":{"passed":true,"evidence":"slogan/premise 来自 description 'Graphs that teach > graphs that impress'；多智能体分析流水线与文件/函数/类/依赖图谱来自 README Features 与 cloud 描述；可搜索/可提问/差异影响来自 'Ask anything about the codebase' 与 'Analyze impact of your current changes'；平台 Claude Code/Codex/Cursor/Copilot/Gemini CLI 来自 description 与 summary claim；Karpathy wiki+力导向社区聚类来自 evidence claim；Figma 模式来自 v2.9.0 release claim；.ua/ 数据目录来自 release claim。均属素材范围内。"},"signatureVisual":{"passed":true,"evidence":"Hero 提供独立 SVG(viewBox 1100x300)：约8度略带弧度的单行轨道(path M40,230 ... 1040,84)，左侧散点文件簇，按序三个发光节点(文件->函数->依赖，间距递增体现收缩凝集并标注 condense/cluster)，右上收敛星群，配色金绿节点+暮紫连接线+青色阶段编号，完全对应 signatureVisual 与 heroGeometry。"},"mobileTransformation":{"passed":true,"evidence":"媒体查询(max-width:719px)将 hero-top 收敛为单列、hero-side 下沉、.cols 变单列、.stages 转为带左边框的纵向阶梯条、foot 竖排，形成真实重构而非仅换色/圆角/模糊；QA 无横向溢出(scrollWidth=390)。"},"localeParity":{"passed":true,"evidence":"zh-CN 与 en 六段结构、章节顺序与角色完全一致；正文为自然译文(如 '扫描一次，之后就能一直追问' / 'Scan once, then keep asking')，量级均可读，无一处遗漏或硬直译；h1/slogan 与默认 locale 契约对应。"},"antiTemplate":{"passed":true,"evidence":"未回退到 Generic Hero/三卡指标墙/卡片墙/设备壳；阶段用错落小标尺、hero 右侧为 Brief 明示的竖排摘要而非指标卡；虽 page::before 含两枚低透明 radial 渐变，但仅作背景氛围，标志视觉为独立轨道 SVG 而非装饰渐变替代，未违反 forbiddenPatterns。"},"noFakeUi":{"passed":true,"evidence":"无伪搜索框/开关/假截图/假测试数据；页面仅含指向 understand-anything.com 与 GitHub 的真实链接与 aria-label，无回调无 JS 假控件；QA fakeControlCount=0。"}},"issues":[{"code":"MISSING_MOBILE_ORBIT_TRANSFORM","severity":"warning","locale":"all","description":"Brief 的 mobileTransformation 明确要求'轨道旋转为纵向阶梯落差，节点自上而下依次点亮，点击任一节点展开对应阶段的简要说明'；实际移动端仅将版式(hero-top/cols/stages/foot)重构为单列，标志轨道仍是未随断点旋转的原 SVG，且节点未提供点击展开交互。","repairInstruction":"在 max-width:719px 断点为 .orbit 的 SVG/节点提供纵向阶梯变换(或改画纵向轨道)，并让三个节点可点击展开对应阶段说明，落实 Brief 的移动端交互规格；若放弃需在文档说明偏差理由。"}]},"repairEvidence":[],"provenance":{"executor":"opencode","model":"deepseek/deepseek-v4-flash","opencodeVersion":"1.1.58","promptVersion":"dodot-showcase-deepseek-v4","inputHash":"52a9b67e1e33f60e3c921978634986f4aae5b6dd4c3f715a49b37e0ab03364a7","artifactHash":"6c3f28305ca852aa0d780c47cc401c8cc306004b99a02c4926c960466a9d261d","designSkill":"dodot-showcase-design"},"qaEvidence":{"zh-CN":[{"viewport":"desktop","passed":true,"localePassed":true,"metrics":{"language":"zh-CN","h1Count":1,"h2Count":4,"sectionCount":4,"textLength":1210,"cjkCount":644,"latinLetterCount":312,"scrollWidth":1280,"clientWidth":1280,"scrollHeight":2804,"h1Top":352.046875,"bottomInset":287.5,"brokenImageCount":0,"missingAltCount":0,"fakeControlCount":0,"emptyLinkCount":0},"screenshotHashes":{"top":"ec3cf69b6c5d27ed600bf1befbeb62b8ab784bfe18c8cb871fc87de54a228b84","middle":"f639dd1a6a634a949ea52cac82fbbb00470db7ba14af52a8e9ba71d256f03029","bottom":"e215fe6e669567cb3776b5751f29a3db3f7039be53ff849de07b7f96d0887d15"},"consoleErrors":[],"networkRequests":[]},{"viewport":"mobile","passed":true,"localePassed":true,"metrics":{"language":"zh-CN","h1Count":1,"h2Count":4,"sectionCount":4,"textLength":1210,"cjkCount":644,"latinLetterCount":312,"scrollWidth":390,"clientWidth":390,"scrollHeight":3844,"h1Top":174.28125,"bottomInset":372.890625,"brokenImageCount":0,"missingAltCount":0,"fakeControlCount":0,"emptyLinkCount":0},"screenshotHashes":{"top":"b509896fb5450197ab1191da576daf19db67ba5a3af939fdbe388bb3e5b8cf4d","middle":"a191c3725b5f3d6d0370a5498e879ef0b76745a2f68436c7badd9194867b54bf","bottom":"c85b08197136422a921de890abc09650d452f4b2f6f00a1f0ec26657853dc6a0"},"consoleErrors":[],"networkRequests":[]}],"en":[{"viewport":"desktop","passed":true,"localePassed":true,"metrics":{"language":"en","h1Count":1,"h2Count":4,"sectionCount":4,"textLength":2882,"cjkCount":0,"latinLetterCount":2267,"scrollWidth":1280,"clientWidth":1280,"scrollHeight":3016,"h1Top":304.375,"bottomInset":287.8125,"brokenImageCount":0,"missingAltCount":0,"fakeControlCount":0,"emptyLinkCount":0},"screenshotHashes":{"top":"6c9c5733a651a098dbeee1f84b8e43191df147729bcabdf3c09e4c393c8336c4","middle":"3bb7d01e1257b895f1ea3ba4d7cabe9800ea25f2c9ad259ff9d0fc83c76b7696","bottom":"29a466f1ec4b37413d4e397bf845f7c1da87b8f700a3e5514281f98e01abefa6"},"consoleErrors":[],"networkRequests":[]},{"viewport":"mobile","passed":true,"localePassed":true,"metrics":{"language":"en","h1Count":1,"h2Count":4,"sectionCount":4,"textLength":2882,"cjkCount":0,"latinLetterCount":2267,"scrollWidth":390,"clientWidth":390,"scrollHeight":4523,"h1Top":174.28125,"bottomInset":373.328125,"brokenImageCount":0,"missingAltCount":0,"fakeControlCount":0,"emptyLinkCount":0},"screenshotHashes":{"top":"9a5e98bc09ed9b4bc3fbf32948fa630303519cc52ecfc838571894ba73b9f64e","middle":"4f6402478df581fcd0c7f831957048876c89f62aee98d0aa103243e2395d3a70","bottom":"e0127767b7c08f5f7bebbcf5163547e4c45cec73f4561acd021186edccd43308"},"consoleErrors":[],"networkRequests":[]}]}},"generatorVersion":"opencode-1.1.58:deepseek/deepseek-v4-flash:dodot-showcase-deepseek-v4","publishedAt":"2026-08-20T10:59:46.214Z"},"rawExtraction":{"projectRevisionId":"cmt2zgzru00go6271n8hebbhu","projectVersion":1,"sources":[{"canonicalKey":"github:egonex-ai/understand-anything","role":"primary","revisionVersion":1,"url":"https://github.com/Egonex-AI/Understand-Anything"}],"facts":[{"key":"repository_owner","value":"egonex-ai","evidence":[{"path":"repository.owner","excerpt":"egonex-ai"}]},{"key":"repository_name","value":"understand-anything","evidence":[{"path":"repository.name","excerpt":"understand-anything"}]},{"key":"description","value":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.","evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]},{"key":"primary_language","value":"TypeScript","evidence":[{"path":"repository.language","excerpt":"TypeScript"}]},{"key":"license","value":"MIT","evidence":[{"path":"repository.license","excerpt":"MIT"}]},{"key":"topics","value":"antigravity-skills, business-knowledge, claude-code, claude-skills, codebase-analysis, codex, codex-skills, developer-tools-ai-agent, gemini-cli-skills, karpathy-llm-wiki, knowledge-base, knowledge-graph, memory, opencode-skills, pi-agent, understandcode, vibe-coding","evidence":[{"path":"repository.topics","excerpt":"antigravity-skills, business-knowledge, claude-code, claude-skills, codebase-analysis, codex, codex-skills, developer-tools-ai-agent, gemini-cli-skills, karpathy-llm-wiki, knowledge-base, knowledge-graph, memory, opencode-skills, pi-agent, understandcode, vibe-coding"}]},{"key":"source_count","value":"1","evidence":[{"path":"projectRevision.sources","excerpt":"github:egonex-ai/understand-anything"}]}],"taxonomy":{"category":"development_tools","subcategory":"data_ai_tools","confidence":0.65,"evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]},"lifecycle":{"stage":"active","confidence":0.75,"evidence":[{"path":"repository.pushedAt","excerpt":"2026-08-11T14:13:37.000Z"}]},"userIntents":[{"slug":"build_software","label":"构建或改进软件","confidence":0.65,"evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]},{"slug":"automate_work","label":"自动化工作流程","confidence":0.65,"evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]},{"slug":"create_content","label":"创作内容或设计","confidence":0.65,"evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]},{"slug":"analyze_data","label":"分析与理解数据","confidence":0.65,"evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]},{"slug":"learn_explore","label":"学习与探索","confidence":0.65,"evidence":[{"path":"repository.description","excerpt":"Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more."}]}],"readme":{"headings":["✨ Features","Explore the structural graph","Understand business logic","Analyze knowledge bases","🚀 Quick Start","1. Install the plugin","2. Analyze your codebase","Generate Chinese content (知识图节点描述和 Dashboard UI)","Supported languages: en (default), zh, zh-TW, ja, ko, ru","3. Explore the dashboard","4. Keep learning","Ask anything about the codebase","Analyze impact of your current changes","Deep-dive into a specific file or function","Generate an onboarding guide for new team members","Extract business domain knowledge (domains, flows, steps)","Analyze a Karpathy-pattern LLM wiki knowledge base","Re-run anytime — incremental by default (only re-analyzes changed files)","Auto-update on every commit via a post-commit hook","Scope to a subdirectory (for huge monorepos)"],"commands":["npx https://github.com/Egonex-AI/Understand-Anything/releases/latest/download/understand-anything-viewer.tgz /path/to/analyzed/project"]},"changeSet":[{"field":"authorName","beforeHash":"74234e98afe7498fb5daf1f36ac2d78acc339464f950703b8c019892f982b90b","afterHash":"fee22b96ffa8a7933959e66bba866bbfd0529c3ccb2df8ed90c29146ea862ce7"},{"field":"description","beforeHash":"74234e98afe7498fb5daf1f36ac2d78acc339464f950703b8c019892f982b90b","afterHash":"2fc39e0293126f26a5c378936d7a459bd0183a820b94a93b33c2b04ffc572cca"},{"field":"experienceUrl","beforeHash":"74234e98afe7498fb5daf1f36ac2d78acc339464f950703b8c019892f982b90b","afterHash":"7d27d562aa6a02ca575acd1ebb7a40b3ef4a85d27a1f44594f7f2de903fdf9a4"},{"field":"name","beforeHash":"74234e98afe7498fb5daf1f36ac2d78acc339464f950703b8c019892f982b90b","afterHash":"1bc0b46402a03e263838a6b249c09610fa1704e0f46957398b7b0ceb952d3c13"},{"field":"repository","beforeHash":"74234e98afe7498fb5daf1f36ac2d78acc339464f950703b8c019892f982b90b","afterHash":"c1e9a35f90922a8cc1038c45a8c9d258d6fee7f73ffbdb6ea15b4802e9706db7"},{"field":"sourceCount","beforeHash":"74234e98afe7498fb5daf1f36ac2d78acc339464f950703b8c019892f982b90b","afterHash":"6b86b273ff34fce19d6b804eff5a3f5747ada4eaa22f1d49c01e52ddb7875b4b"}]},"publishedAt":"2026-08-10T10:55:56.551Z","updatedAt":"2026-08-21T13:26:49.662Z","summary":{"id":"cmt2zgzry00gq6271qb6clval","projectType":"agent","oneLine":"将代码库、知识库或文档转化为交互式知识图谱","overview":"Understand Anything 是一个 AI 编码助手插件，通过多智能体流水线分析代码库、知识库或文档，构建包含文件、函数、类和依赖的知识图谱，并提供可搜索、可提问的交互式仪表盘。支持 Claude Code、Codex、Cursor、Copilot、Gemini CLI 等平台，帮助开发者快速理解大型代码库。","highlights":["多智能体分析流水线","交互式知识图谱仪表盘","支持多种 AI 编码平台","Figma 设计图分析"],"bestFor":["需要快速上手大型代码库的开发者","希望可视化理解项目架构的团队","使用 AI 编码工具的用户"],"caveats":["初始分析大型项目可能消耗较多 token","部分功能依赖 LLM 生成，效果可能因模型而异","证据基于 README 和发布说明，未验证实际效果"],"categorySuggestion":"development_tools","subcategorySuggestion":"coding_agents","lifecycleStages":["idea_requirements","frontend","backend_data","testing_quality"],"audienceIntents":["discover","learn","build","evaluate"],"confidence":0.85,"evidence":[{"claim":"Understand Anything 是一个 Claude Code 插件，通过多智能体流水线分析项目，构建包含文件、函数、类和依赖的知识图谱，并提供交互式仪表盘。","sourceType":"readme","locator":"https://github.com/Egonex-AI/Understand-Anything"},{"claim":"支持多种平台，包括 Claude Code、Codex、Cursor、Copilot、Gemini CLI 等，可通过安装脚本或插件市场安装。","sourceType":"readme","locator":"https://github.com/Egonex-AI/Understand-Anything"},{"claim":"提供引导式导览、模糊与语义搜索、差异影响分析、角色自适应 UI、分层可视化和语言概念解释等功能。","sourceType":"readme","locator":"https://github.com/Egonex-AI/Understand-Anything"},{"claim":"支持分析 Karpathy 模式的 LLM wiki 知识库，生成带社区聚类的力导向知识图谱。","sourceType":"readme","locator":"https://github.com/Egonex-AI/Understand-Anything"},{"claim":"v2.9.0 新增 Figma 设计图分析模式，可将 Figma 文件转换为可导航的知识图谱，并支持设计系统模型。","sourceType":"release","locator":"https://github.com/Egonex-AI/Understand-Anything/releases/tag/v2.9.0"},{"claim":"数据目录已重命名为 .ua/，并保持向后兼容，旧版本项目无需迁移。","sourceType":"release","locator":"https://github.com/Egonex-AI/Understand-Anything/releases/tag/v2.9.0"}],"publishedAt":"2026-08-17T13:58:54.095Z"},"github":{"url":"https://github.com/Egonex-AI/Understand-Anything","ownerAvatarUrl":"https://avatars.githubusercontent.com/u/257477979?v=4","starsCount":79593,"forksCount":6683,"language":"TypeScript","syncedAt":"2026-08-17T13:55:30.452Z","license":"MIT","defaultBranch":"main","archived":false,"disabled":false,"lastChangedAt":"2026-08-11T14:13:37.000Z","missing":false},"tags":[{"name":"Vibe coding","slug":"vibe-coding","verification":"verified","evidenceUrl":"https://github.com/Egonex-AI/Understand-Anything","evidenceExcerpt":"GitHub topic: vibe-coding"}],"updates":[]},"entryPoint":"ranking","claim":{"state":"unclaimed","verified":false},"initialView":"beautified","community":{"viewer":null,"state":{"likesCount":0,"creativeCount":0,"followersCount":0,"commentsCount":0,"liked":false,"creative":false,"followed":false},"comments":{"comments":[],"nextCursor":null}}}]}],"$L24"]