ai-infrastructure-hugg…
Hugging Face Inference SDK patterns for TypeScript/Node.js — InferenceClient setup, chat completion, text generation, streaming, embeddings, image generation,…
VitePress 1.x — Vue-powered static site generator for documentation sites, built on Vite
$ npx -y skills add agents-inc/skills --skill web-meta-framework-vitepress --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/web-meta-framework-vitepressContext preview
The summary Claude sees to decide when to auto-load this skill.
VitePress 1.x — Vue-powered static site generator for documentation sites, built on Vite
name: web-meta-framework-vitepress description: VitePress 1.x — Vue-powered static site generator for documentation sites, built on Vite
> **Quick Guide:** VitePress is a Vue-powered static site generator for documentation, configured > entirely in `.vitepress/config.ts`. The sidebar is either an array (one sidebar everywhere) or an > object keyed by URL path prefix (a sidebar per section). Data loaders — files ending `.data.ts` — run > at build time and ship only their serialized result to the client. Vue components work directly inside > Markdown through `<script setup>`. Every page is pre-rendered at build time, which is the constraint > behind most of the red flags below. > > **Version:** VitePress 1.6.x, on Vite 6+ and Vue 3.5+.
**Detailed Resources:**
---
<critical_requirements>
**Put site configuration in `.vitepress/config.ts` and wrap it in `defineConfig()`.** That path is where the build looks, and the wrapper is what gives the option names type checking.
**Load data through a `*.data.ts` loader rather than fetching in a component.** The loader runs at build time and the client receives only the serialized result, which is both faster and SSR-safe.
**Guard browser APIs with `<ClientOnly>` or `onMounted`.** Every page is pre-rendered at build time, so a bare `window` or `document` reference crashes the build rather than the browser.
**Extend the default theme with `extends: DefaultTheme` plus layout slots.** A fork is a copy that stops tracking upstream, and the slot list is broad enough that forking is rarely the shorter route.
**Build Markdown collection pages with `createContentLoader()`.** It already handles the glob, the frontmatter extraction, mtime caching and dev-mode watching.
</critical_requirements>
---
**Auto-detection:** VitePress, vitepress, .vitepress/config, defineConfig vitepress, createContentLoader, vitepress/theme, DefaultTheme, useData, useSidebar, markdown-it plugin vitepress, vitepress deploy
**Applies to:**
**Handled elsewhere:**
---
<patterns>
One file, wrapped for type checking. `cleanUrls` drops `.html` from URLs, `sitemap.hostname` generates `sitemap.xml`, `lastUpdated` reads git timestamps, and `search.provider: "local"` is search with no service to sign up for.
import { defineConfig } from "vitepress";
export default defineConfig({
title: "My Docs",
cleanUrls: true,
lastUpdated: true,
sitemap: { hostname: "https://docs.example.com" },
themeConfig: {
nav: [{ text: "Guide", link: "/guide/" }],
sidebar: {
/* Pattern 2 */
},
search: { provider: "local" },
editLink: { pattern: "https://github.com/org/repo/edit/main/docs/:path" },
},
});Full code: [examples/core.md](examples/core.md)
An array gives one sidebar for the whole site. An object keyed by path prefix gives a different sidebar per section, and the first matching prefix wins.
sidebar: {
"/guide/": [
{ text: "Getting Started", collapsed: false, items: [{ text: "Introduction", link: "/guide/introduction" }] },
{ text: "Advanced", collapsed: true, items: [{ text: "Data Loaders", link: "/guide/data-loading" }] },
],
"/api/": [{ text: "API Reference", items: [{ text: "Config", link: "/api/config" }] }],
}The trailing slash matters: `/guide` also matches `/guidelines`. Omitting `collapsed` makes a group permanently expanded rather than collapsible.
Full code: [examples/core.md](examples/core.md)
A `*.data.ts` file runs at build time and exports `data`. `createContentLoader` covers Markdown collections; a plain object with `watch` and `load()` covers everything else.
// posts.data.ts
import { createContentLoader } from "vitepress";
export default createContentLoader("blog/posts/*.md", {
excerpt: true,
transform: (raw) =>
raw
.sort(
(a, b) => +new Date(b.frontmatter.date) - +new Date(a.frontmatter.date),
)
.map(({ url, frontmatter, excerpt }) => ({
title: frontmatter.title,
url,
excerpt,
})),
});<script setup>
import { data as posts } from "./posts.data";
</script>`transform` is where you d
The official skills marketplace for Agents Inc. 150+ skills covering everything from React and Prisma to Redis, ElevenLabs, and infrastructure tooling. Pick the skills that match your stack and install them via Claude Code. Need more control?
Repo: agents-inc/skills
Hugging Face Inference SDK patterns for TypeScript/Node.js — InferenceClient setup, chat completion, text generation, streaming, embeddings, image generation,…
LiteLLM proxy server setup, TypeScript client patterns via OpenAI SDK, model routing, fallbacks, load balancing, spend tracking, virtual keys, and production…
Serverless GPU compute platform for AI model deployment — web endpoints, GPU functions, model serving, and TypeScript client patterns
Local LLM inference with the Ollama JavaScript client -- chat, streaming, tool calling, vision, embeddings, structured output, model management, and…
Replicate SDK patterns for TypeScript/Node.js -- client setup, predictions, streaming, webhooks, file handling, model versioning, deployments, and training
Together AI SDK patterns for TypeScript — client setup, chat completions, streaming, structured output, function calling, embeddings, image generation,…