convez.ai
MCP ServerDesign & UIIntermediário
ddruids

Mobbin Skill - Pesquisa Estruturada de UI/UX

ddruids/mobbin-skill

Converte pesquisa de UI/UX em padrões de design estruturados via Mobbin

Ver no GitHub
BomPluginScore · de 0 a 100

MCP Server que transforma a biblioteca de 300K+ telas do Mobbin em pesquisa de UI estruturada. Analisa visualmente padrões de design, fluxos de usuário e componentes de sistemas de design, gerando insights agrupados por padrões recorrentes. Funciona com Claude Code, Cursor, Gemini CLI, Lovable e qualquer agente compatível com MCP.

Instalação

$npx convez-ai ddruids/mobbin-skill

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
40
Forks
9
Acompanhando
0
Saúde
42%
Último commit
há 1 mês
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Pesquisar padrões de onboarding em apps fintech
  • Comparar fluxos de confirmação de transação entre wallets de cripto
  • Extrair componentes de design system para dashboards SaaS
  • Analisar tratamento de estados vazios em apps iOS
  • Derivar queries de pesquisa a partir de PRDs

Veredito

Ferramenta bem estruturada que automatiza análise visual de UI patterns em escala. Requer conta Mobbin e setup inicial, mas entrega síntese de pesquisa com qualidade profissional para decisões de design baseadas em dados.

README

A tradução para o português deste README ainda está na fila.

Mobbin UI Research Skill

Turn Mobbin's 300K+ screen library into structured UI research. Give your agent a research goal — it searches Mobbin for screens and flows, visually analyzes the results, and returns pattern clusters, design system components, or competitive comparisons.

Works with Claude Code, Cursor, Gemini CLI, Lovable, and any MCP-compatible agent.

What it does

  • Translates abstract UX goals into concrete Mobbin search queries
  • Searches both individual screens and multi-step user flows (onboarding, checkout, etc.)
  • Visually analyzes returned screenshots (layout, hierarchy, components, color)
  • Clusters findings into recurring UI patterns with linked references
  • Derives research queries from PRDs and product briefs automatically
  • Surfaces atomic design system primitives when asked about components
  • Supports three output modes: research summary, competitive comparison, product decision log

Install

Requires Mobbin MCP to be connected first. Copy and Paste in Terminal to install skill

npx skills add https://github.com/ddruids/mobbin-skill

Usage

Once installed, the skill activates when you ask about UI research, screen patterns, or design system components:

"Research how fintech apps handle onboarding"
"Compare how crypto wallets display transaction confirmation"
"What are common components for a SaaS dashboard design system?"
"How do AI apps handle empty states on iOS?"
"Here's my PRD — research the key screens and flows"

The skill automatically:

  1. Resolves the target platform (iOS or web)
  2. Picks the right tool mix — search_screens for specific screens, search_flows for multi-step journeys
  3. Generates 3-7 concrete search queries (or derives them from your PRD)
  4. Analyzes the returned screenshots and flow sequences
  5. Synthesizes findings into structured output with Mobbin URLs

What's inside

FilePurpose
SKILL.mdCore skill — screen/flow tool selection, query construction, PRD-driven research, analysis framework, output formats, design system component mode
references/query-patterns.mdQuery formula, platform-specific guidance, 20+ screen query examples, flow query examples by category
references/synthesis-framework.md8 analysis lenses, 3 output templates (research summary, competitive comparison, product decision log)

  1. Sign up at mobbin.com
  2. Click your profile icon (top right) → Settings → MCP
  3. Select your tool and follow the instructions, or use one of these:

Claude Code:

claude mcp add mobbin \
  --transport http https://api.mobbin.com/mcp

Cursor:

{
  "mcpServers": {
    "mobbin": {
      "serverUrl": "https://api.mobbin.com/mcp"
    }
  }
}

Codex:

codex mcp add mobbin --url https://api.mobbin.com/mcp

Lovable:

https://api.mobbin.com/mcp

License

MIT

Ver o original no GitHub

Bom para

Relacionados