alchaincyf/huashu-design — explained in plain English
Analysis updated 2026-06-24
Build a clickable mobile app prototype with multiple navigating screens just by describing the app to your AI agent.
Generate a browser slide deck that also exports as a real editable PowerPoint file, not an image.
Create a print-quality infographic by describing the content and optionally supplying brand assets.
Get an expert critique of any existing design scored on five dimensions with a prioritized improvement list.
| alchaincyf/huashu-design | mmistakes/minimal-mistakes | adityatelange/hugo-papermod | |
|---|---|---|---|
| Stars | 13,554 | 13,493 | 13,485 |
| Language | HTML | HTML | HTML |
| Setup difficulty | easy | easy | easy |
| Complexity | 2/5 | 2/5 | 2/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Internal agent prompts are in Chinese but the skill understands and responds to English tasks.
Huashu Design is a design skill you install into an AI coding agent, like Claude Code or Cursor, that lets you describe what you want in plain language and receive a finished visual output in minutes. You run one command to add it, then type instructions to your agent the same way you would talk to a designer. The agent does the rest: it picks styles, pulls in assets, writes the code, and delivers something you can actually use. The skill covers several distinct output types. You can get a clickable mobile app prototype with a realistic iPhone frame and multiple screens that actually navigate between each other. You can get a slide deck that works in the browser and also exports as a real PowerPoint file with editable text, not an image. You can get an animation or short promotional video rendered out as an MP4 or GIF. You can get a print-quality infographic with precise typography and real data. Each of these takes roughly 3 to 30 minutes depending on complexity. The built-in design system contains 20 distinct design philosophies, so even if you provide no brand assets, the output avoids the generic look common to AI-generated visuals. If you do have a brand, the skill walks through a five-step process to find your logo, colors, fonts, and screenshots before generating anything, so the result matches your actual brand rather than a guess. There is also a five-dimension expert critique mode. You can pass any design to it and receive scores on philosophical coherence, visual hierarchy, execution craft, functionality, and innovation, along with a prioritized list of what to keep, what to fix, and quick improvements you can make right away. The skill is agent-agnostic, meaning it works with Claude Code, Cursor, Trae, and other tools that support markdown-based skills. It was built by a Chinese-speaking developer and the internal agent prompts are in Chinese, but the skill understands and responds to English tasks without issue. The license is personal use only.
Huashu Design is a skill for AI coding agents like Claude Code or Cursor that turns plain-English descriptions into finished prototypes, slide decks, animations, and infographics in minutes, with a built-in expert critique mode.
Mainly HTML. The stack also includes HTML, JavaScript.
Personal use only, commercial use is not permitted.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.