| name | brand-voice |
| description | Define brand voice, tone, vocabulary, Do/Don't rules, and AI Brand Kit. Use when: codifying how the brand speaks for consistent content creation. Two modes: Quick (Layer 0, ~30min, from URL or 5 questions) produces Voice Snapshot; Full (Layer 4, ~2-3h, after ECPs + positioning) produces complete Voice Guide + AI Brand Kit + Per-ECP and Per-Channel adaptations. Quick mode MANDATORY before any content. Full mode adds ECP tone shifts, channel guidance, vocabulary rules, Do/Don't library, and 5-check Voice Test. NOT for: visual identity (use visual-identity), content creation (use content skills), or positioning strategy (use positioning-messaging). |
| metadata | {"author":"Alfonso Sainz de Baranda (Growth4U)","version":"3.0","system":"SanchoCMO","phase":"1","pillar":"brand-voice","layer":"0+4","depends_on":"none (Quick) | niche-discovery-100x, positioning-messaging (Full)","updated":"2026-02-27","changes":"v3 — Restructured per skill-creator principles. SKILL.md lean (~130 lines). Concepts/questions in references."} |
| context_required | ["brand/{slug}/company-brief/company-brief.current.md","brand/{slug}/go-to-market/positioning/*/*.current.md","brand/{slug}/company-brief/company-brief.current.md"] |
| context_writes | ["brand/{slug}/brand-voice/brand-voice.current.md"] |
Brand Voice
Cómo habla la brand — codificado para consistencia. Quick Snapshot al inicio; Full Voice Guide + AI Brand Kit cuando ECPs y positioning estén listos.
Input: URL/materiales (Quick) | ECPs + positioning + content samples (Full)
Output: Voice Snapshot / AI Brand Kit → brand/{slug}/brand-voice/brand-voice.current.md
References
| Archivo | Cuándo leer | Contenido |
|---|
| hydration.md | SIEMPRE — Step 0 obligatorio | Mapeo de campos upstream → esta skill |
| prompt.md | SIEMPRE — fuente de verdad | 6 pattern categories (Extract) + 15 Strategic Questions (Build) + Build Pipeline |
| checklist.md | Antes de entregar — self-QA | Ítems Quick + Full mode |
| concepts.md | Si necesitas modos, AI Brand Kit spec, Voice Test | Definiciones, edge cases, Lite/Deep |
| schema.md | Si necesitas el schema de campos | Estructura de datos completa |
Flujo de Ejecución
0. Context Hydration (OBLIGATORIO — antes de cualquier pregunta)
- Lee
_system/skills/context-hydration-protocol.md para el patrón genérico
- Lee
references/hydration.md para el mapeo específico de esta skill
- Lee TODOS los docs en
context_required
- Pre-rellena campos según hydration_map
- Presenta datos heredados al usuario: "De [fuente] ya tengo X. ¿Correcto?"
- Solo pregunta campos listados en "Campos genuinamente nuevos"
Quick Mode (Layer 0 — siempre primero)
- Elegir path: URL disponible → Path A (scrape + extract) | Sin URL → Path B (5 Quick Questions)
- Analizar: Patterns de voz (tono, vocabulario, POV, ritmo) + visual notes ligeras
- Generar Voice Snapshot: 3 adjectives + tone spectrum + words USE/AVOID + Do/Don't pairs + examples by type + confidence level
- Presentar al usuario para validación
Full Mode (Layer 4 — cuando ECPs + positioning listos)
- Cargar inputs: Voice Snapshot + ECPs seleccionados + Messaging Playbook + Positioning data + 3-5 content samples
- Elegir approach según confidence del Snapshot: Extract (high) | Hybrid (medium) | Build (low)
- Extract: Analizar 6 pattern categories (ver
references/prompt.md)
Build: Usar 15 Strategic Questions + Build Pipeline (ver references/prompt.md)
- Generar Voice Profile completo: traits, tone spectrum (5 dims), vocabulary, rhythm, boundaries
- Per-ECP Adaptation: tone shift + vocabulary + proof emphasis + channels por cada ECP
- Per-Channel Guidance: tone flex + length + structure + example por canal activo
- Generar AI Brand Kit: 2-3 páginas AI-loadable (Voice DNA + Do/Don't + Vocab + Rhythm + Channel ref + ECP cheat sheet)
- Voice Test: 5 checks (recognizable, actionable, differentiated, authentic, consistent). Iterar 1-2 rondas.
Self-QA (OBLIGATORIO)
- Lee
references/checklist.md
- 0 ❌ antes de entregar
- Metadata:
<!-- Self-QA: PASS | fecha | items: X✅ Y⚠️ 0❌ -->
Guardar con versionado
- Ruta:
brand/{slug}/brand-voice/brand-voice.current.md
- Backup + versionado + history.json
- Link:
<MC_BASE>/docs/brand/{slug}/brand-voice/brand-voice.current.md
Cross-Pillar Data Flow
| Dato | Lo consume |
|---|
| Voice Snapshot (3 adj + tone) | ALL content skills |
| AI Brand Kit | ALL downstream — loaded before any content generation |
| Words USE/AVOID | Every piece of copy |
| Do/Don't library | Content quality checks, AI prompts |
| Per-ECP tone adaptation | positioning-messaging, outreach-workflow |
| Per-channel guidance | social-content, email-sequences, paid-ads |
| Light Visual Notes | visual-identity (input for deep visual work) |
🔬 Profundizar con Deep Research
Al entregar, añade:
📊 **¿Quieres profundizar?**
→ Escribe **"profundizar"** para continuar.
📁 Almacenamiento (OBLIGATORIO)
brand/{{slug}}/brand-voice/
├── brand-voice.current.md ← versión activa
├── v1.md, v2.md... ← versiones anteriores
├── history.json ← log de versiones
└── qa-log.md ← historial de QA
- Identifica slug desde systemPrompt
- Si existe
brand-voice.current.md → backup como v{N+1}.md
- Si no existe → crea carpeta +
brand-voice.current.md + v1.md + history.json
- Link:
<MC_BASE>/docs/brand/{slug}/brand-voice/brand-voice.current.md