Communitygithub.com

majiayu000/claude-skill-registry

Use when user wants to explore, compare, or A/B test multiple design styles, visual directions, or mockups for a web page before implementation begins. Triggers on "show me design options", "style exploration", "design comparison", "visual options".

What is claude-skill-registry?

claude-skill-registry is a Claude Code agent skill that use when user wants to explore, compare, or A/B test multiple design styles, visual directions, or mockups for a web page before implementation begins. Triggers on "show me design options", "style exploration", "design comparison", "visual options".

Works with✓Claude Code~Codex CLI~Cursor
npx skills add https://github.com/majiayu000/claude-skill-registry/tree/HEAD/skills/design/web-design-wizard

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Web Design Wizard

Overview

Generate multiple distinct design styles for a web page in parallel worktrees, letting the user compare and refine before committing to a direction. Core principle: explore divergently, then converge on the best option.

When to Use

  • User wants to explore different design styles for a web page
  • User asks for design options, visual direction, or style comparison
  • User wants multiple implementations to choose from

Do NOT use when:

  • User has a specific design already defined → use saurun:react-tailwind-v4-components or frontend-design directly
  • User wants to modify an existing design, not create new ones

Prerequisites

All skills below must be installed. If any are missing, stop and inform the user.

SkillUsed InPurpose
ui-ux-pro-maxPhase 2Design system generation
superpowers:using-git-worktreesPhase 6Parallel worktree setup
frontend-designPhase 6Production-grade implementation
vercel-react-best-practicesPhase 6React optimization
copywritingPhase 6Copy refinement
nano-banana-proPhase 6Image generation — install: npx skills add intellectronica/agent-skills@nano-banana-pro -g -y

Workflow

digraph design_wizard {
    "User request" [shape=doublecircle];
    "Discovery" [shape=box];
    "Style Generation" [shape=box];
    "User satisfied?" [shape=diamond];
    "Refinement" [shape=box];
    "Graphics Preference" [shape=box];
    "Project Setup" [shape=box];
    "Parallel Implementation" [shape=box];
    "Parallel Verification" [shape=box];
    "Complete" [shape=doublecircle];

    "User request" -> "Discovery";
    "Discovery" -> "Style Generation";
    "Style Generation" -> "User satisfied?";
    "User satisfied?" -> "Refinement" [label="no"];
    "Refinement" -> "User satisfied?";
    "User satisfied?" -> "Graphics Preference" [label="yes"];
    "Graphics Preference" -> "Project Setup";
    "Project Setup" -> "Parallel Implementation";
    "Parallel Implementation" -> "Parallel Verification";
    "Parallel Verification" -> "Complete";
}

Phase 1: Discovery

Use AskUserQuestion to gather context. Batch questions — use up to 4 questions per call. Infer obvious answers from context (e.g., a coffee shop is consumer-facing, probably warm mood).

Gather:

  1. Project name — detect from package.json or git repo name, confirm with user
  2. Page type — Landing, Portfolio, Blog, E-commerce, SaaS, Documentation
  3. Target audience — Professionals, Consumers, Developers, Creative industry
  4. Mood/feeling — Professional, Playful, Luxury, Minimal, Bold, Warm, Futuristic
  5. Content language — English, Danish, etc.
  6. Style preferences/anti-preferences — only if not obvious from above

Skip questions where the answer is obvious from the user's request.

Phase 2: Style Generation

Generate design styles using ui-ux-pro-max:

Default: 4 styles (user can request different count).

For each style:

  1. Craft diverse keywords from discovery responses. Combine: page type + audience + mood + unique aesthetic. Ensure variety across all styles.

  2. Generate design system: Invoke ui-ux-pro-max via the Skill tool with arguments: "[keywords]" --design-system --persist -p "[ProjectName]" --page "[style-name]" Saves to: design-system/pages/[style-name].md in current working directory.

  3. Present to user with: style name, key colors, typography, mood adjectives, and best use case.

Track generated styles to avoid duplicate keywords in refinement rounds.

Phase 3: Refinement (Iterative)

User SaysAction
"Give me N completely different styles"Generate N NEW styles with different keywords
"Give me N more options"Generate N NEW (avoid previous keywords)
"I like X and Y, give me more like X"Keep X & Y, generate variations of X
"Combine styles X and Y"Generate 1 new blending those aesthetics
"Implement X, Y, Z" or "Implement all"Move to Phase 4 with selections

Phase 4: Graphics Preference

Before implementation, ask about graphics for all selected styles at once:

Use AskUserQuestion with options:

  • "Lots of graphics for all (hero + accent graphics + illustrations)"
  • "Minimal graphics for all (hero image only)"
  • "Let the design style determine graphics" (Recommended)
  • "I want to choose per style"

Only ask per-style if user picks the last option.

Phase 5: Project Setup

Confirm project location

Suggested: designs/[project-name]/ in current working directory.

Ask user to confirm or override.

Create project

  1. mkdir -p [location]
  2. Write concept.md — see concept-template.md for template. Populate from discovery responses. Only fill fields you have data for — omit sections with no info.
  3. Do NOT run git init — use the existing repo. Commit project docs to the current branch.
git add designs/[project-name]/concept.md
git commit -m "docs: initialize [project-name] design exploration"

Project structure

[repo-root]/
├── designs/[project-name]/
│   └── concept.md
├── design-system/pages/
│   ├── [style-1].md
│   └── [style-2].md
└── .worktrees/
    ├── [style-1-branch]/    # React + Vite + TypeScript project
    └── [style-2-branch]/

Phase 6: Parallel Implementation

CRITICAL: Spawn ALL implementation agents in a SINGLE message with multiple Task tool calls.

For each selected style, spawn one sub-agent. See implementation-prompt.md for the full prompt template. Replace [absolute-path-to-repo-root] with the actual repo root path so sub-agents can resolve cross-worktree references.

Key points for each sub-agent:

  • Use superpowers:using-git-worktrees to create worktree at .worktrees/[style-name]/
  • Create React + Vite + TypeScript project: npm create vite@latest . -- --template react-ts
  • Read design spec from design-system/pages/[style-name].md
  • Use Skill tool to invoke: frontend-design, vercel-react-best-practices, copywriting, nano-banana-pro
  • Generate images per graphics preference
  • Build with npm run build to verify
  • Commit all changes

Wait for ALL agents to complete before Phase 7.

Phase 7: Parallel Verification

CRITICAL: Spawn ALL verification agents in a SINGLE message with multiple Task tool calls.

For each implementation, spawn one verification sub-agent. See verification-prompt.md for the full prompt template. Assign ports sequentially (5173, 5174, 5175, ...) and replace [absolute-path-to-repo-root] with the actual path.

Verification checks:

  • Dev server starts without errors
  • All URLs/links are valid
  • Images exist and are properly referenced
  • Design spec adherence (color, typography, layout, effects) — score 1-10
  • Production build succeeds

After all complete, show comparison summary:

Style 1: [Name] — ✅ All passed — Design: 8.5/10
Style 2: [Name] — ⚠️ 2 broken URLs — Design: 9.0/10
Style 3: [Name] — ❌ Build failed — See report

Recommend best-performing style based on results.

Quick Reference

PhaseActionKey Tool
1. DiscoveryGather context via AskUserQuestion (batch up to 4)AskUserQuestion
2. Style GenerationGenerate 4 design systems via ui-ux-pro-maxSkill
3. RefinementIterate until user selects stylesAskUserQuestion
4. GraphicsAsk graphics preference for selected stylesAskUserQuestion
5. Project SetupCreate concept.md, commit scaffoldingWrite, Bash
6. ImplementationSpawn parallel sub-agents per styleTask
7. VerificationSpawn parallel verification agents, show comparisonTask

Common Mistakes

  • Regenerating design specs in sub-agents — design specs are created in Phase 2. Sub-agents READ them, never re-invoke ui-ux-pro-max.
  • Sequential sub-agent spawning — Phase 6 and 7 MUST spawn ALL agents in a single message with multiple Task calls. Never one-at-a-time.
  • Skipping Graphics Preference — always ask before implementation. Don't assume.
  • Using git add . in worktrees — stage specific paths only to avoid committing node_modules, .env, or OS files.
  • Identical keywords across styles — track used keywords to ensure diversity. Duplicate keywords produce duplicate designs.

Error Handling

ErrorRecovery
ui-ux-pro-max script failsRetry with simpler keywords. If persistent, create design system manually from the style keywords.
All styles rejectedAsk user for specific direction: "What did you dislike? Colors? Layout? Overall feel?"
Sub-agent skill not foundThe agent should use Skill tool to invoke. If skill genuinely missing, skip that enhancement step and note it.
Implementation build failsFix in worktree, re-run build. Report issue if unfixable.

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

majiayu000/claude-skill-registry

Builds, edit or validate Agent Skill through conversational discovery. Use when the user requests to "Create an Agent", "Optimize an Agent" or "Edit an Agent".

majiayu000/claude-skill-registry

Edit existing BMAD agents while maintaining compliance

majiayu000/claude-skill-registry

Edit existing BMAD agents while maintaining compliance

majiayu000/claude-skill-registry

Edit existing BMAD modules while maintaining coherence

majiayu000/claude-skill-registry

Edit existing BMAD modules while maintaining coherence

majiayu000/claude-skill-registry

Analyzes current state and user query to answer BMad questions or recommend the next workflow or agent. Use when user says what should I do next, what do I do now, or asks a question about BMad

majiayu000/claude-skill-registry

Build AI agents with Google ADK Python (Agent Development Kit). Use for multi-agent systems, workflow agents (sequential/parallel/loop), Vertex AI deployment, tool integration, human-in-the-loop.

majiayu000/claude-skill-registry

Build AI agents with Google ADK Python (Agent Development Kit). Use for multi-agent systems, workflow agents (sequential/parallel/loop), Vertex AI deployment, tool integration, human-in-the-loop.

majiayu000/claude-skill-registry

Portfolio allocation and rebalancing optimizer. Manages asset allocation across stocks/cash/bonds, performs periodic rebalancing, and ensures diversification according to market regime and risk tolerance.

majiayu000/claude-skill-registry

Create a hilarious and ultra-realistic video of an anthropomorphic animal acting like a human vlogger in a real-world setting.

majiayu000/claude-skill-registry

Speech-to-text transcription and translation via OpenAI Audio API -- models, response formats, timestamps, prompting, streaming, chunking, and diarization

majiayu000/claude-skill-registry

BlockRun works with Claude Code and Google Antigravity.

majiayu000/claude-skill-registry

This skill should be used when the user asks about libraries, frameworks, API references, or needs code examples. Activates for setup questions, code generation involving libraries, or mentions of specific frameworks like React, Vue, Next.js, Prisma, Supabase, etc.

majiayu000/claude-skill-registry

小省导购员数字人带货版即梦视频提示词生成系统,基于四大智能体协同(提示词生成师、质量管控师、知识库运维师、跨环节适配师),按照"主体+运动+场景+(镜头语言+光影+氛围)"公式输出中英文双版提示词,适配5s短视频。确保人物一致性、视觉连贯性、情绪连贯性,支持知识库智能复用和跨工具适配(Suno音乐、AI绘画),为数字人带货视频提供高质量提示词生成服务。

majiayu000/claude-skill-registry

Upload and manage files using Google Gemini File API via scripts/. Use for uploading images, audio, video, PDFs, and other files for use with Gemini models. Supports file upload, status checking, and file management. Triggers on "upload file", "file API", "upload image", "upload PDF", "upload video", "file management".

majiayu000/claude-skill-registry

CRITICAL: Use for MolyKit AI chat toolkit. Triggers on: BotClient, OpenAI, SSE streaming, AI chat, molykit, PlatformSend, spawn(), ThreadToken, cross-platform async, Chat widget, Messages, PromptInput, Avatar, LLM

majiayu000/claude-skill-registry

AI and machine learning workflow covering LLM application development, RAG implementation, agent architecture, ML pipelines, and AI-powered features.

majiayu000/claude-skill-registry

Analyze images/audio/video with Gemini API (better vision than Claude). Generate images (Imagen 4), videos (Veo 3). Use for vision analysis, transcription, OCR, design extraction, multimodal AI.

majiayu000/claude-skill-registry

Analyze images/audio/video with Gemini API (better vision than Claude). Generate images (Imagen 4), videos (Veo 3). Use for vision analysis, transcription, OCR, design extraction, multimodal AI.

majiayu000/claude-skill-registry

Skill for discovering and researching autonomous AI agents, tools, and ecosystems using the AgentFolio directory.

Related Skills