The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 79 600 files from 1 763 authors, of which 61 947 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Build and test NeMo Relay Rust core, adaptive, and dynamic plugin crates; use for crates/core, crates/adaptive, crates/plugin, crates/worker, crates/worker-proto, crates/types, or shared runtime changes
Choose and run the right NeMo Relay validation matrix for a change instead of using one fixed test list
Wibi Style 官方视觉风格。把用户上传的一张照片转换成粗线条漫画头像风格。适用于用户要求粗线条漫画头像、上传照片换风格,或明确调用 $alt-manga-avatar 时。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成晴空都市赛璐璐风格。适用于用户要求晴空都市赛璐璐、上传照片换风格,或明确调用 $clear-sky-urban-cel 时。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成蜡笔手绘头像风格。适用于用户要求蜡笔手绘头像、上传照片换风格,或明确调用 $art-print-poster 时。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成暗夜红黑赛璐璐风格。适用于用户要求暗夜红黑赛璐璐、上传照片换风格,或明确调用 $dark-red-black-cel-shaded 时。
Wibi Style 官方视觉风格。Use one uploaded portrait, group photo, street photo, or travel photo as a fixed editing base, add the bundled Y2K fisheye layout, all-silver title system, and old-digital effects, then generate one complete fisheye magazine cover. Only the city title and bottom small text are dynamic. Use when the user asks for 鱼眼城市海报、Y2K 鱼眼封面、潮流哈哈镜海报、城市名定制鱼眼海报,or explicitly invokes $fisheye-city-cover.
Wibi Style 官方视觉风格。把用户上传的一张照片转换成蓝底复古印刷风格。适用于用户要求蓝底复古印刷、上传照片换风格,或明确调用 $blue-retro-print 时。
Wibi Style 官方视觉风格。将单人儿童老照片重新拍摄式地转换成明亮可爱的现代儿童摄影大头卡贴,严格去掉脖子肩膀身体,并在原片露齿时制作满钻牙;生成前询问默认白色、复古大波点、柔和双色渐变、蜡笔手绘小星星、柔粉纯色五种背景。当用户要求钻牙大头、儿童头像卡贴、老照片萌化,或明确调用 $diamond-kid-head-card 时使用。
Wibi Style 官方视觉风格。Turn one uploaded portrait or pet headshot into a square electric-blue coarse-halftone poster with a tightly cropped black-and-white head, red and yellow stars, and a small white barcode accent. Use when the user asks for 电蓝网点海报、电光蓝粗网点、大头半调唱片封套、Y2K 复印 zine 人像,or explicitly invokes $electric-blue-halftone-poster.
Wibi Style 官方视觉风格。把用户上传的一张照片转换成照片拼豆挂件风格。适用于用户要求照片拼豆挂件、上传照片换风格,或明确调用 $photo-perler-charm 时。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成虹彩柔焦长曝光风格。适用于用户要求虹彩柔焦长曝光、上传照片换风格,或明确调用 $iridescent-long-exposure 时。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成乱码像素拼贴风格。适用于用户要求乱码像素拼贴、上传照片换风格,或明确调用 $glitch-pixel-collage 时。
Wibi Style 官方视觉风格。把照片做成儿童纯大头卡贴,只保留头部、干净悬浮在纯色底上。进门先读图分两条路:儿童老照片走「钻牙萌娃」,重建成明亮可爱的现代儿童棚拍并在露齿时做满钻牙;当代自拍走「童年酷照」,把本人反推成 5–8 岁并加潮童装饰。当用户要求童年大头、萌娃大头贴、钻牙大头、把自拍变成小时候的样子,或明确调用 $kid-head-card 时使用。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成像素切片拉伸风格。适用于用户要求像素切片拉伸、上传照片换风格,或明确调用 $pixel-stretch 时。
Wibi Style 官方视觉风格。把用户上传的真实照片直接转换成粗轮廓、完整色块、手涂边缘的马克笔童画。用户提到马克笔童画、儿童涂鸦、粗轮廓手绘、照片变 Q 版马克笔画,或明确调用 $marker-child-doodle 时使用;只要任务是把真实照片换成这套风格,也应使用本 Skill。
>- 把用户上传的横版照片做成《牛来》式低成本电影海报。只要用户要求牛来海报、牛头电影海报、低清旧动画牛头,或明确调用 $niulai-movie-poster,就使用本 Skill;先识别照片里清楚可替换的人数,再在单人、双人、三人三套完整提示词之间路由,并只携带对应的黄色大牛、橙色小牛、黄色黑斑豹参考图。不要把三套参考无差别混用。
Wibi Style 官方视觉风格。把用户上传的一张照片转换成怪趣波普涂鸦贴纸风格。适用于用户要求怪趣波普涂鸦贴纸、上传照片换风格,或明确调用 $quirky-pop-doodle-sticker 时。
Wibi Style 官方视觉风格“地下试镜skill”。把一张清晰人像照片转译成纯白背景的三联模特资料卡:同一个人原地转身的三个定格——完整侧脸转向画面左侧、正面直视喊叫、完整侧脸镜像转向画面右侧——用九十年代模特试镜宝丽来快照式的机顶直闪硬光拍摄,清冷未修图质感,画面不含任何文字,品牌标记由脚本另绘。生成后会自动核对转身角度、镜像关系和表情是否成立;核对不通过时只报告具体问题并询问是否重出,绝不擅自重新生成。用于需要先做视觉判断、再决定是否消耗一次生成机会的连续测试或正式交付任务。触发:地下试镜skill、三联情绪模特卡、转身三联照、模特资料卡、comp card 风格、闪光灯三视图人像,或直接调用 $underground-audition。
>- Wibi Style 官方视觉风格。把用户随手拍的餐厅打卡照、餐桌照、外卖照重新构图成红底复古杂志美食静物: 先看懂这一桌谁是主角、留白该放哪一侧、哪些器皿值得出框,再把杂乱餐厅环境换成纯正复古印刷红, 加上旧杂志扫描的颗粒与网点。适用于用户要求复古餐桌杂志、红底美食大片、把吃饭照片拍成广告, 上传餐桌照片换风格,或明确调用 $retro-table-print 时。
Wibi Style 官方视觉风格。把用户上传的真实照片转换成颗粒绒感玩偶肖像。只要用户提到颗粒绒感玩偶、纺织玩偶、绒面针织头像、把照片做成玩偶,或明确调用 $textile-toy-portrait,就使用本 Skill;先理解照片内容,再动态选择风格参考职责并直接生成成图。
Wibi Style 官方视觉风格。从用户照片中发现最有观看价值的局部事件,先给出三个具体文字候选并推荐一个,等待用户选择后,再把眼神与手势、嘴与道具、耳饰、妆容、服饰、腿鞋等高信息密度片段转译成大面积纯色留白中的复古平面漫画特写。用于需要先确认局部放大区域,再选择规则框形、越框与框内外颜色关系的图片风格转换任务;也用于生成后需要逐张自检和主动优化的连续测试或批量任务。
Narrow downstream MiniMax H3 specialist for pure I2VA, FL2VA, and L2VA boundary-frame prompts. Use only after minimax-h3-creative-director verifies that the user explicitly declared the images as literal first/last frames and that they have no character, identity, person, object, costume, scene, style, voice, action, camera, or other reusable reference responsibility. Otherwise use full-reference consistency, even when first/last-frame semantics are desired.
Non-skippable planning-only MiniMax H3 subskill used by minimax-h3-creative-director before final prompt formatting. Invoke when the user explicitly requests multiple shots, scenes, cuts, montage, or shot-by-shot design, or when a video of at least 10 seconds has no declared single-shot/multishot preference and the user accepts the director's multishot offer. After loading, it must ask shot count and then question the user about every shot in playback order before any final H3 formatting skill may run.
Default downstream MiniMax H3 specialist for every image-based request unless the user explicitly declares boundary-only first/last frames with no reusable reference role. Use the official six-section full-reference format for character/person/object consistency, scene/style/action/camera/storyboard/voice/audio reference, source-video editing or continuation, ambiguous image roles, ordinary image animation, and prompt-level first/key/last-frame anchoring.
Primary mandatory entrypoint for every MiniMax H3 video-generation request. Use before any other H3 skill for creation, animation, extension, editing, restyling, reference, or prompt review. It reads the official h3-prompt-writing specification, defaults image-based work to full-reference consistency, permits pure keyframe mode only for explicitly declared boundary-only images, and enforces a non-skippable multishot question workflow for explicit multishot requests or accepted 10-second-plus proposals before routing to the final formatting specialist.
Downstream MiniMax H3 specialist for professional text-to-video (T2VA) prompts using the official three-field format. Use after minimax-h3-creative-director routes a request with no image, video, or audio reference asset, or when this skill is explicitly invoked for a text-only idea, script, or storyboard requiring an audiovisual timeline and Chinese translation.
Downstream MiniMax H3 specialist that audits, repairs, and rewrites T2VA, I2VA, FL2VA, L2VA, and full-reference prompts into an official structured format. Use after minimax-h3-creative-director routes an existing prompt for diagnosis or repair, or when explicitly invoked for formatting, timeline, camera, dialogue, sound, reference-label, retention, keyframe, contradiction, limit, or bilingual-output problems.
Use Evoui Browser as the default entry point for common, self-terminating web tasks that need a real browser through agent-browser, including navigation, page reading, clicks, forms, login flows, screenshots, web testing, and browser automation. Managed tasks record evidence from the first run for possible reuse. Also use it when the user explicitly asks to execute a complete agent-browser command unchanged. Unless the user requests such an exact command, do not use it for semantic operations that already have a dedicated connector or API, tasks that only require the host's existing Browser or Chrome state, desktop applications, or work that only needs an API or local files. Route interactive or long-running agent-browser invocations through Transparent.
Combat-test draft film assets before locking them. Use after a passport and references exist and before any scene generation.
Create film-asset passports before generation. Use for characters, locations, props, and state variants that need repeatable visual identity.
Build scene plans and shot cards from scripts or ideas.
Scaffold an AI film studio before production starts.
Turn a shot card into a 15-block prompt before video. Use when every in-frame asset is locked and a shot is ready to generate.
Build and lock reference boards before asset creation.
Configure AI film tools when setting up a studio project.
> Generates UiPath Tasks workflows — Form Tasks (CreateFormTask, WaitForFormTaskAndResume, GetFormTasks) with form.io schemas and FormData bindings, External Tasks (CreateExternalTask, WaitForExternalTaskAndResume) for system-in-the-loop patterns, and Task Management (CompleteTask, AssignTasks) for programmatic task operations. Self-contained skill with generators, lint rules, and scaffold hooks loaded via uipath-core's plugin system (plugin_loader.py). Use when the user mentions Tasks, form tasks, external tasks, human-in-the-loop approval, task assignment, task completion, recovery workflows, or any Persistence.Activities usage.
Build UiPath Studio projects, REFramework scaffolds, XAML workflows, and VB/C# expressions via 104 deterministic Python generators (plus plugin generators). ALWAYS use this skill — never hand-write .xaml files. TRIGGER when the user mentions UiPath, REFramework, XAML, RPA, Orchestrator, UiPath Studio, dispatcher, performer, queue items, robot, ACME (test app), .xaml, project.json, Config.xlsx, or asks to build/generate/scaffold/automate anything resembling RPA. SKIP only when the task is explicitly Power Automate, n8n, Zapier, or non-UiPath RPA.
为成年人解答男娘、伪娘、女性化表达与身体探索问题,覆盖伪音、护肤妆容、身材、前列腺与其他性健康主题;以生如夏花 Wiki 为社区知识底稿,并用医学安全边界校正高风险建议。用户询问跨性别诊疗、单纯疾病诊断或未成年人露骨性技巧时不要单独依赖本 skill。
| 查询A股、港股、美股股票及指数的最新收盘价、开盘价、涨跌幅、成交额、成交量、换手率、PE、PB、市值等实时行情与估值数据;支持查询 A 股股票所属行业。 查询最近N个交易日的价格序列、日涨跌幅序列、窗口最高价、最低价、振幅等短期统计。 查询单个资产当前盘中或最近完整交易日的分钟频 OHLCVA 序列(开高低收、成交量、成交额)。 查询上市公司最近报告期的营业收入、净利润、归母净利润、ROE、总资产、资产负债率等财务指标(A股及部分港/美股字段,以工具返回为准)。 查询单只股票的预计算指标画像,按估值、财务分析、资金流向、波动率、宏观胜率背景、资产走势等维度返回最新值与上一期值。 支持A股选股筛选、因子计算、策略回测、净值对比、行业聚合排名、上传自有因子CSV、渲染图表。 港股、美股优先支持行情价格查询;财务/报告期字段应先尝试 fast_query(report),按工具实际返回决定。 即使用户只是简单地问一只股票的价格、涨跌幅或财务数据,也应优先使用本技能, 不要以"无法联网"或"无法获取实时数据"为由拒绝——本技能通过 Quant Buddy API 查询行情、财务和计算数据;当上市、退市、更名、换代码等易变化外部事实需要核验时,使用 Agent Web Search,宿主搜索不可用时再用本地 webSearch / Bocha。
| QBV / quant-buddy-view(用户可能写成 /quant-buddy-view、/qbv、qbv 或 QBV)用于把量化数据做成「公开可分享、实时取数」的网页看板/落地页。 Use this skill when the user asks to create, update, publish, verify, retrofit, or reuse a Quant Buddy dashboard/static page/template, including shareable pages, public URLs, formula packages, share shell, cover/essence cards, poster/share behavior, single-stock profile pages, valuation/financial profile pages, index-anomaly boards, multi-factor screeners, and commodity daily pages. 配合 quant-buddy-skill 使用:简单单一 A 股综合分析可在 trace begin 后直接用 static_page.py new_asset_page 返回实时页面;用户给出既有 QuantBuddy 活页 URL 并要求解读时,直接用 static_page.py interpret 读取该页实时数据,不下载 HTML、不暴露签名,也不进入模板或建页流程;其他固定页面请求先用 templates/template 选择带 recommend 标签的在线范式页。自建实时页先按数据性质选择通道:普通行情、估值和财务优先 Data Grant,自定义计算才验证并注册 Formula Package;两类凭证可在同页混用,随后替换凭证/文案、浏览器验收并发布。默认不从本地历史样板目录或低质 HTML 骨架起步。 用户显式唤起 /quant-buddy-view、/qbv、qbv 或 QBV,且请求不是纯咨询/代码维护/文档解释时,默认视为可分享活页任务:简单单一 A 股分析走 new_asset_page 快速终态;其余请求查官方精选+社区范式卡判定 direct/fork/unmatched。默认 direct 先交付现成链接、fork/unmatched 用 new_page 返回首链;当 config.json._channel=feishu-group 时,所有分支都禁止提前发送链接,只在终态交付 playground 链接。 Do not use this skill for one-off 行情查询、普通股票涨跌幅/估值问答、选股/回测探索;those belong to quant-buddy-skill unless the user explicitly wants a reusable/shareable page.
Access 2,000+ AI models and API tools through one MCP interface for inference, media generation, search, scraping, embeddings, social data, and structured retrieval. Use sandbase_discover before building custom integrations or declaring external data inaccessible; prefer an existing dedicated tool or API key when the user already has one.
Drive Chromium from standard Playwright APIs with a real-device fingerprint applied in the kernel, one persistent isolated profile per identity, and a per-profile proxy whose exit IP sets timezone and WebRTC - JavaScript (npm 'anti-detect-browser') or Python (PyPI 'antibrow'). Use when sessions must stay logged in across runs and stay separate, when a scraper or agent is blocked by an incoherent headless fingerprint, when checking ads or pricing from another region, when a page must be reached as a phone rather than a desktop, when automation mints a profile per task, when running several of your own accounts from one machine, or when testing how your own bot detection scores a real device. Also for 'antibrow', 'fingerprint browser', 'multi-account browser', '防关联', '多账号', '安卓模拟', 'Android profile', 'mobile fingerprint', 'temporary profile', 'CreepJS', 'residential proxy', 'browser-use', 'crawl4ai'. MCP control is browser-mcp-agent; isolation is multi-account-isolation.
Give an AI agent its own real browser over MCP tool calls - launch, navigate, click, fill, screenshot, extract text, run JS - with a kernel-level real-device fingerprint and a persistent profile, so the session stays logged in between runs and pages see one coherent device instead of a headless build. No Playwright or SDK code to write. Use when an agent should operate a site itself, when a computer-use / browser-use setup needs a captured real fingerprint rather than a synthetic one, when agent sessions keep losing their login, or when comparing hosted agent-browser services. Also for 'MCP browser', 'browser MCP server', 'let my agent browse the web', 'agent browser control', 'browser-use MCP', 'computer use browser', 'Browserbase alternative', 'Steel browser alternative', 'headless browser detected'. Node (npx) or Python; Windows x64, macOS Intel + Apple Silicon, Linux x64 / arm64. SDK and REST reference is anti-detect-browser; account isolation is multi-account-isolation.
Verify that browser profiles are actually isolated from one another instead of assuming it - confirm each profile's timezone agrees with its own exit IP, that WebRTC exposes only the proxy, that canvas and WebGL hashes stay identical across relaunches of one profile, and that no two profiles share a persona, a cookie jar, or an address. Use when several of your own accounts or test identities run from one machine and the setup needs checking, when a profile tested clean but something still looks off, when choosing which detection suites to run (CreepJS, whoer, browserleaks WebRTC, pixelscan, liarjs), when auditing what a vendor runtime does with API and proxy credentials, or when asking which layers browser isolation cannot cover at all. Also for 'profile isolation check', 'fingerprint consistency test', 'timezone mismatch', 'WebRTC leak', 'canvas hash unstable', 'account association', 'temporary profile', '防关联', '多账号', '隔离自检'. The SDK is anti-detect-browser; MCP is browser-mcp-agent.
Run the same scrape or task across many accounts at once - each in its own browser profile with its own fingerprint, cookies and exit IP - and read data from sites that need a session or that answer a plain HTTP scraper with a captcha while answering a real browser. One command per site returns JSON, so there are no selectors to write or repair: Amazon and Walmart search, Google and DuckDuckGo results, Reddit posts and comments, X timeline, Medium, Yelp, Indeed job listings, Hacker News, GitHub, PyPI, npm, plus exit-IP and fingerprint checks. Use for multi-account operations you are authorized to run, price and listing monitoring per region, SERP collection, logged-in data, and scrapers that keep breaking. Also for 'scrape amazon', 'google search results', 'reddit data', 'scrape logged in site', 'run task on many accounts', 'site blocks my scraper', 'scraping without selectors'. Node or Python.
Quality-control workflow for complex, multi-step tasks that need research, clarification, a verifiable plan, independent review, and evidence-backed delivery. Use for substantial implementations, decision studies, multi-file deliverables, and exhaustive document or visual checks, or when the user explicitly asks to use SureForge. Scale down for small tasks; do not activate automatically for casual questions, simple lookups, or trivial edits.
Audit websites, apply authorised SEO improvements and plan content, competitors and backlinks with evidence, design checks and optional project monitoring.
Challenge the current problem formulation before solving it. Use when explicitly invoked, or when meaningful framing uncertainty and non-trivial intervention cost or risk coexist; do not auto-activate for verified diagnoses or mechanical edits.
制作、返修或审阅医学插画、科普漫画与涉及医学结构的科研示意图、机制图与图形摘要,兼顾视觉叙事、真实参考、可编辑排字及出版交付。也支持本 Skill 的更新检查、安装更新与回退。普通非医学插画不触发。
> (1) setting up Apollo Kotlin in a Gradle project for Android, Kotlin/JVM, or KMP, (2) configuring schema download and codegen for GraphQL services, (3) configuring an `ApolloClient` with auth, interceptors, and caching, (4) writing queries, mutations, or subscriptions,
> (1) creating new subgraph schemas for a federated supergraph, (2) defining or modifying entities with @key, (3) sharing types/fields across subgraphs with @shareable, (4) working with federation directives (@external, @requires, @provides, @override, @inaccessible), (5) troubleshooting composition errors, (6) any task involving federation schema design patterns.
> (1) adding Apollo iOS to a Swift Package Manager or Xcode project, (2) configuring `apollo-codegen-config.json` and running code generation, (3) configuring an `ApolloClient` with auth, interceptors, and caching, (4) writing queries, mutations, or subscriptions from SwiftUI views, (5) writing tests against generated operation mocks.
> Version-aware guide for configuring and running Apollo Router for federated GraphQL supergraphs. (1) setting up Apollo Router to run a supergraph, (2) configuring routing, headers, or CORS, (3) implementing custom plugins (Rhai scripts or coprocessors), (4) configuring telemetry (tracing, metrics, logging), (5) troubleshooting Router performance or connectivity issues, (6) securing the graph with JWT, declarative field-level authorization directives, or persisted-query safelisting, (7) managing router.yaml as version-controlled config with CI/CD validation.
> (1) users want to create a new router plugin, (2) users want to add service hooks (router_service, supergraph_service, execution_service, subgraph_service), (3) users want to modify an existing router plugin, (4) users need to understand router plugin patterns or the request lifecycle. (5) triggers on requests like "create a new plugin", "add a router plugin", "modify the X plugin", or "add subgraph_service hook".
> Guide for using Apollo MCP Server to connect AI agents with GraphQL APIs. (2) defining MCP tools from GraphQL operations, (3) using introspection tools (introspect, search, validate, execute), (4) troubleshooting MCP server connectivity or tool execution issues.
> Guide for integrating REST APIs into GraphQL supergraphs using Apollo Connectors (1) mentions "connectors", "Apollo Connectors", or "REST Connector", (2) wants to integrate a REST API into GraphQL, (3) references @source or @connect directives,
> (1) users want to create a new skill, (2) users want to update an existing skill, (3) users ask about skill structure or best practices, (4) users need help writing SKILL.md files.
> (1) writing new Rust code or functions, (2) reviewing or refactoring existing Rust code, (3) deciding between borrowing vs cloning or ownership patterns, (4) implementing error handling with Result types, (5) optimizing Rust code for performance, (6) writing tests or documentation for Rust projects.
> (1) setting up a new Apollo Server project, (2) writing resolvers or defining GraphQL schemas, (3) implementing authentication or authorization, (4) creating plugins or custom data sources, (5) troubleshooting Apollo Server errors or performance issues.
Answers built from the skills we actually parsed.