菜单
Análisis y generación de voz de marca, tono e identidad comunicacional. Usar cuando: "voz de marca", "tono de marca", "brand guidelines", "branding", "rebranding", "cómo suena mi marca", "estilo de comunicación", "quiero definir el tono de mi empresa". Genera BRAND-VOICE.md.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Vercel a profundidad — CLI avanzado, CI/CD, env vars y patrones Next.js App Router (Next 16). Reemplaza al plugin oficial de Vercel (43 skills) condensado a lo esencial. Complementa a `deployment` (que cubre el deploy básico). Usar cuando: "vercel link", "monorepo vercel", "vercel build --prebuilt", "vercel promote", "vercel rollback", "CI/CD vercel", "GitHub Actions vercel", "VERCEL_TOKEN", "vercel env pull", "OIDC token", "VERCEL_OIDC_TOKEN", "project.json vs repo.json", "App Router", "server component async", "cookies() headers() params async", "middleware proxy.ts", "revalidateTag", "next/router next/navigation", "getServerSideProps migrar", "next 16 breaking changes". Do NOT use for: deploy básico a Vercel (usar `deployment`), Railway, Express puro, ni AI SDK/Workflow/shadcn del plugin (esos quedaron fuera del recorte — reactivar el plugin oficial si se necesitan).
Autenticación en Next.js: NextAuth v5 (Auth.js), JWT, sesiones, middleware y RBAC. Stack: Next.js App Router + NextAuth v5 / Supabase Auth / JWT custom. Usar cuando: "NextAuth", "Auth.js", "login", "sesión", "JWT", "middleware de auth", "proteger rutas", "RBAC", "roles y permisos", "OAuth", "credentials provider", "useSession", "getServerSession", "auth middleware", "redirect si no autenticado".
Colas de trabajo asíncronas con BullMQ + Redis en Node.js/Express. Reintentos, concurrencia, delays anti-ban, workers y monitoreo. Usar cuando: "cola de mensajes", "job queue", "BullMQ", "tareas en background", "envío masivo", "anti-ban WhatsApp", "rate limiting mensajes", "worker", "reintentar tarea fallida", "procesar en segundo plano", "Bull", "Redis queue", "scheduled jobs", "delayed jobs".
Agrega un chatbot, asistente IA o chat widget a un sitio web o app. Usar cuando: "chatbot", "chat widget", "asistente virtual", "chat flotante", "bot de ventas", "bot de soporte", "FAQ bot", "useChat", "streaming chat", "Vercel AI SDK", "asistente inteligente", "chat con IA", "embedded chat".
Integra un CMS headless a Next.js para que el cliente edite contenido sin código. Stack: Sanity / Contentful / Strapi. Usar cuando: "CMS", "headless CMS", "Sanity", "sitio administrable", "blog con CMS", "el cliente quiere editar", "GROQ", "live preview", "contenido dinámico", "Strapi".
Dashboards y visualización de datos en React/Next.js. Stack: Recharts / Chart.js. Usar cuando: "dashboard", "gráfica", "chart", "KPIs", "métricas", "stat cards", "bar chart", "pie chart", "line chart", "heatmap", "panel de analytics", "reporting dashboard", "admin panel con datos", "visualizar datos".
| name | market-brand |
| description | Análisis y generación de voz de marca, tono e identidad comunicacional. Usar cuando: "voz de marca", "tono de marca", "brand guidelines", "branding", "rebranding", "cómo suena mi marca", "estilo de comunicación", "quiero definir el tono de mi empresa". Genera BRAND-VOICE.md. |
You are the brand intelligence engine for /market brand <url>. Analyze a brand's voice across all available channels and generate a complete brand voice guidelines document. Output to BRAND-VOICE.md.
For dimension scoring tables, archetype descriptions, tone shift maps, vocabulary frameworks, competitor voice matrices, and consistency audit templates, load: references/brand-voice-frameworks.md
Formato de salida: ver references/output-template.md.
Analyze multiple sources — a voice analysis from one page is incomplete.
Primary (analyze first): Homepage, About page, Product/service pages
Secondary: Blog posts (3-5 recent), Social media (bio + recent posts), Email newsletters, Customer-facing microcopy (errors, onboarding, help docs)
Tertiary: Job postings, Press releases, Ad copy, Video/podcast transcripts
Use WebFetch for URL-based sources. Find social profile links from the target site.
Map brand on four spectrums (1-10). Load references/brand-voice-frameworks.md for signal tables.
For each: state score, quote 3-5 examples from source material, explain writing implications. Evidence is non-negotiable.
Map to one of five archetypes. Load references/brand-voice-frameworks.md for descriptions and key phrases:
State: primary archetype, secondary (if applicable), fit strength (Strong/Moderate/Weak), evidence.
Voice = consistent personality. Tone shifts by context. Load references/brand-voice-frameworks.md for the tone shift map template. Fill with real quotes from each context.
Extract linguistic fingerprint. Load references/brand-voice-frameworks.md for framework.
Compare against 2-3 competitors. Load references/brand-voice-frameworks.md for the Voice Comparison Matrix. Identify: differentiated territory, overlap (risk), unclaimed positioning.
Score consistency across all channels. Load references/brand-voice-frameworks.md for the scorecard. Frame inconsistencies as opportunities.
| Level | Format | Length |
|---|---|---|
| Tagline | Under 10 words | 6-9 words |
| Value propositions | 1 sentence each | 3-5 props |
| Elevator pitch | Conversational | 75 words |
| Boilerplate | Press / bios | 100-150 words |
| Brand story | Full narrative | 300-500 words |
For each: document current state, assess alignment with identified voice, draft improved version where needed.
Generate "Our Voice IS / IS NOT" chart, then specific Do's and Don'ts. Make each instruction concrete enough for a freelancer to apply without follow-up questions (e.g., "use contractions" not "be friendly").
Generate 8 samples — people learn voice by example:
COMPETITOR-REPORT.md exists, use competitor data for voice comparisonCOPY-SUGGESTIONS.md exists, cross-reference value propositions/market copy to rewrite copy in identified voice, /market competitors for deeper competitive voice analysis