web▌

Expert 3D web implementation across Three.js, React Three Fiber, Spline, and WebGL with performance optimization. \n \n Covers three primary stacks (Spline for rapid prototyping, React Three Fiber for React apps, Three.js vanilla for maximum control) with decision guidance on when to use each \n Includes 3D model pipeline: format selection, poly-count reduction, texture baking, GLB export, and compression with gltf-transform \n Handles scroll-driven 3D interactions using ScrollControls, GSAP, an

Orchestrates multi-source web research by delegating to subagents, synthesizing findings, and producing cited reports. \n \n Breaks research questions into 2–5 distinct subtopics, creates a research plan file, and spawns up to 3 subagents in parallel for efficient investigation \n Each subagent performs 3–5 web searches per subtopic and writes findings to local files with key facts, quotes, and source URLs \n Synthesizes results by reading local findings files, integrating insights across subtop

Extract clean Markdown article content from URLs with three-tier fallback strategies. \n \n Implements cascading extraction methods: Jina Reader (fast, 200 requests/day free), Scrapling + html2text (unlimited, handles paywalled content), and direct web_fetch (static pages fallback) \n Preserves Markdown structure including headings, links, images, lists, code blocks, and blockquotes \n Domain-aware routing skips Jina for WeChat articles, Zhihu, Juejin, and CSDN to conserve quota and improve succ

Web search with ranked results, snippets, and support for freshness filters, SafeSearch, and custom ranking via Goggles. \n \n Returns structured results across multiple types: web pages, news, videos, discussions, FAQ, infobox, and locations in a single response \n Supports freshness filtering (past day/week/month/year or custom date ranges), SafeSearch levels, and location-aware results via headers \n Goggles enable custom result ranking: boost trusted sources, suppress spam, or build focused

Systematic approach to measuring, analyzing, and improving website loading speed and Core Web Vitals. \n \n Covers five-step optimization workflow: establish baseline metrics with Lighthouse, identify bottlenecks (bundle size, images, render-blocking resources), prioritize high-impact fixes, implement improvements, and verify results \n Addresses Core Web Vitals (LCP, FID, CLS) with specific code examples for image optimization, code splitting, lazy loading, and layout shift prevention \n Includ

Measures Core Web Vitals and identifies render-blocking resources, network chains, layout shifts, and caching issues. \n \n Analyzes performance traces using Chrome DevTools MCP to extract FCP, LCP, TBT, CLS, Speed Index, and TTFB with detailed breakdowns \n Identifies render-blocking CSS/JS, network dependency chains, missing preloads, and caching header gaps through network request inspection \n Detects layout shift culprits (images without dimensions, injected content, font swaps) and accessi

Web page content extraction with automatic title, HTML, and metadata retrieval. \n \n Fetches and parses web pages via the page_reader function, returning structured data including title, HTML content, plain text, and publication time \n Supports both CLI usage for quick tasks and SDK integration for programmatic access in backend code only \n Includes caching, rate limiting, parallel processing, and error handling patterns for production applications \n Handles multiple URLs, content aggregatio

Comprehensive quality review covering performance, accessibility, SEO, and best practices across 150+ checks. \n \n Analyzes code and projects across four audit categories: Performance (Core Web Vitals, resource optimization, loading strategy), Accessibility (perceivable, operable, understandable, robust), SEO (crawlability, on-page, technical), and Best Practices (security, standards, UX patterns) \n Categorizes findings by severity (Critical, High, Medium, Low) with specific, actionable recomm

Audit UI code against Web Interface Guidelines with live rule fetching. \n \n Fetches the latest guidelines from a remote source before each review, ensuring compliance checks stay current \n Accepts file paths or patterns as arguments; prompts for input if none provided \n Outputs findings in concise file:line format for quick scanning and remediation \n Covers accessibility, design patterns, and best practices as defined in the fetched ruleset \n

To build powerful frontend claude.ai artifacts, follow these steps: