define-brand-guidelines

What this skill does

# Brand Guidelines Skill

Create the BRAND_GUIDELINES.md document that defines HOW to communicate with your customer. This document drives all copy and messaging: app UI, marketing, support, emails, everything.

> **Prerequisite**: CUSTOMER.md must exist. Brand guidelines without customer definition is just aesthetic preference. The voice must resonate with WHO you're talking to.

## Overview

This skill supports both **creating new brand guidelines** and **refining existing ones**.

This skill guides you through:
0. **Prerequisite Check** - Verify CUSTOMER.md exists; stop if not
1. **Discovery** - Questions about brand personality, voice, language preferences
2. **Draft Generation** - Create BRAND_GUIDELINES.md based on inputs
3. **Refinement** - Test with sample copy, iterate until voice feels right

## Workflow

### Phase 0: Prerequisite Check

**CRITICAL**: Before anything else, check for CUSTOMER.md:

1. Use Glob to search for `**/CUSTOMER.md` in the current directory
2. **If NOT found**: Stop immediately and inform the user:

```
"I can't create brand guidelines without knowing WHO you're talking to.

Please create your CUSTOMER.md first using /define-customer.

Brand voice without customer definition is just aesthetic preference - it won't resonate with anyone specific."
```

Do NOT proceed. End the workflow here.

3. **If found**: Read the CUSTOMER.md and extract key context:
   - ICP definition (who they are)
   - Pain points (what problems they have)
   - What they value (in a solution)
   - Anti-personas (who they're NOT)
   - Any language/communication hints

**IMPORTANT - Pre-fill Recommendations**: Use CUSTOMER.md to infer recommended options for all questions. Examples:

| If CUSTOMER.md says... | Recommend... |
|------------------------|--------------|
| ICP values "speed", "efficiency", "no patience" | Direct communication, short copy |
| ICP is technical (developers, engineers) | Technical language okay, precision matters |
| ICP values "data", "statistics", "proof" | Data-driven persuasion style |
| Anti-persona is "purists" or "academics" | Avoid being preachy or condescending |
| ICP is "fun-seekers", "casual players" | More playful tone, casual formality |
| ICP is "executives", "professionals" | More formal, authoritative personality |

The goal: **User should be able to accept all recommended defaults** and get a solid brand guide. Only ask them to deviate where CUSTOMER.md doesn't provide clear signals.

Then check for existing BRAND_GUIDELINES.md:

```
header: "Existing Brand Guidelines Found"
question: "I found existing BRAND_GUIDELINES.md. What would you like to do?"
options:
  - "Refine it - update based on new insights"
  - "Start fresh - create new brand guidelines"
  - "Review it - just read through what's there"
```

### Phase 1: Discovery

Use AskUserQuestion for all questions. **Put the recommended option FIRST** with "(Recommended)" suffix. Infer recommendations from CUSTOMER.md.

**Question 1: Brand Personality**

Infer from CUSTOMER.md:
- Data-driven ICP → "Authoritative expert"
- Beginners/learners → "Friendly mentor"
- Contrarian/challengers → "Provocative challenger"
- Enterprise/professional → "Calm professional"
- Fun-seekers/enthusiasts → "Energetic enthusiast" or "Witty companion"

```
header: "Brand Personality"
question: "If your brand was a person speaking to your customer, who would they be?"
options:
  - "[Inferred from CUSTOMER.md] (Recommended)"
  - "Authoritative expert - confident, definitive, data-driven"
  - "Friendly mentor - approachable, helpful, encouraging"
  - "Provocative challenger - bold, contrarian, challenges assumptions"
  - "Calm professional - measured, trustworthy, understated"
  - "Energetic enthusiast - excited, passionate, motivating"
  - "Witty companion - clever, playful, personality-forward"
```

**Question 2: Voice Dimensions**

Infer each dimension from CUSTOMER.md. Put recommended first.

```
header: "Formality"
question: "How formal is your brand's voice?"
options:
  - "[Inferred] (Recommended)"  # e.g., "Casual" if ICP is blitz players
  - "Very formal - professional, polished, no contractions"
  - "Somewhat formal - professional but approachable"
  - "Neutral - depends on context"
  - "Casual - relaxed, contractions okay, conversational"
  - "Very casual - informal, slang okay, like texting a friend"
```

Inference rules for formality:
- Enterprise/executives → Very formal or Somewhat formal
- Developers → Casual or Neutral
- Consumers/players → Casual or Very casual
- Default → Casual (most approachable)

```
header: "Tone Weight"
question: "How serious vs playful is your brand?"
options:
  - "[Inferred] (Recommended)"  # e.g., "Mostly serious" if data-driven
  - "Very serious - no humor, all business"
  - "Mostly serious - occasional lightness"
  - "Balanced - serious when needed, light when appropriate"
  - "Mostly playful - humor is part of the brand"
  - "Very playful - fun and entertainment are core"
```

Inference rules for tone:
- ICP values "fun", "enjoyment" → Mostly playful or Balanced
- ICP values "data", "precision" → Mostly serious
- B2B/professional → Balanced or Mostly serious
- Default → Balanced

```
header: "Technical Level"
question: "How technical is your language?"
options:
  - "[Inferred] (Recommended)"
  - "Highly technical - jargon expected, precision matters"
  - "Somewhat technical - domain terms with explanation"
  - "Accessible - simple language, avoid jargon"
  - "Very simple - anyone should understand"
```

Inference rules:
- ICP is developers/engineers → Highly technical or Somewhat technical
- ICP values "plain language", "accessible" → Accessible
- General consumers → Very simple or Accessible
- Default → Somewhat technical

```
header: "Directness"
question: "How direct is your communication?"
options:
  - "[Inferred] (Recommended)"
  - "Very direct - commands, no hedging, get to the point"
  - "Direct - clear and straightforward"
  - "Balanced - direct but diplomatic"
  - "Soft - suggestive, options-focused"
  - "Very soft - gentle, lots of qualifiers"
```

Inference rules:
- ICP values "speed", "efficiency", "no patience" → Very direct
- ICP is time-constrained → Very direct or Direct
- ICP is beginners/learners → Balanced or Soft
- Default → Direct

**Question 3: Writing Style**

```
header: "Copy Style"
question: "What does your ideal copy look like?"
options:
  - "[Inferred options pre-selected] (Recommended)"
  - "Short and punchy - minimal words, maximum impact"
  - "Concise but complete - efficient, no fluff"
  - "Conversational flow - natural, like talking"
  - "Rich and detailed - thorough explanations"
multiSelect: true
```

Inference: If ICP values speed → "Short and punchy". If ICP is technical → "Concise but complete".

```
header: "Persuasion Style"
question: "How do you persuade?"
options:
  - "[Inferred options pre-selected] (Recommended)"
  - "Data and evidence - stats, proof, numbers"
  - "Benefits and outcomes - what they'll achieve"
  - "Emotional resonance - how they'll feel"
  - "Social proof - others trust us"
  - "Authority - we're the experts"
multiSelect: true
```

Inference: Match to "What the ICP Values" from CUSTOMER.md.

**Question 4: Language Preferences**

```
header: "Language Rules"
question: "Select your language preferences:"
options:
  - "[Inferred bundle] (Recommended)"  # Pre-select compatible options
  - "Use contractions (we're, you'll, it's)"
  - "Avoid contractions (we are, you will, it is)"
  - "Emoji okay in appropriate contexts"
  - "No emoji ever"
  - "Exclamation marks okay (sparingly)"
  - "No exclamation marks"
  - "Industry jargon okay for our audience"
  - "Avoid all jargon"
multiSelect: true
```

Default recommendation: "Use contractions" + "No emoji" + "Industry jargon okay" (professional but approachable)

**Question 5: Anti-Patterns**

```
header: "Voice Anti-Patterns"
question: "What should your brand NEVER sound like?"
options:
  - "[Inferred from anti-personas] (Recommende