/brand-voice -- Brand Voice Engine
Generic copy converts worse than copy with a distinct voice. Not because the
words are different -- because the reader feels like they're hearing from a
PERSON, not a marketing team.
This skill defines that voice. Either by extracting it from existing content,
building it strategically from scratch, or auto-scraping a URL to analyze a
brand's public presence.
For output formatting, see /cmo rules/output-format.md.
On Activation
- Read
brand/ context:
- positioning.md — If available, use market angles to inform voice differentiation. A brand positioned as "the anti-course course" needs a different voice than one positioned as "enterprise-grade."
- audience.md — If available, tailor vocabulary and tone to the target reader. Technical builders get different language than busy executives.
- If neither exists, proceed with defaults. Ask the user about their market and audience during the question sequence.
- Check if
./brand/voice-profile.md already exists:
- Exists → Update mode: show current profile summary, ask what to change.
- Does not exist → Proceed to Mode Selection below.
- Assess the user's request to determine which mode to use (Extract, Build, or Auto-Scrape).
Three Modes
Mode 1: Extract (analyze existing content)
User provides 3-5 pieces of existing content (blog posts, emails, social posts, about page, etc).
The analysis process:
- Read all provided content -- look for patterns, not individual word choices
- Identify the voice DNA:
- Sentence structure (short/punchy vs. flowing/complex)
- Vocabulary level (everyday vs. technical vs. academic)
- Emotional register (warm/friendly vs. authoritative vs. irreverent)
- Perspective (first-person storytelling vs. objective analysis)
- Rhythm (varies sentence length? staccato? musical?)
- Map the personality axes (see Voice DNA Framework below)
- Generate the voice profile
- Write to ./brand/voice-profile.md
Mode 2: Build (construct from scratch)
No existing content. Build voice strategically through guided questions.
The question sequence:
- "If your brand were a person at a dinner party, how would they talk?"
- "Who is your audience? Technical builders, creative professionals, busy executives, everyday consumers, or a mix?"
- "Name 2-3 brands whose tone you admire (not their product -- their VOICE)."
- "What words should NEVER appear in your marketing? What feels off-brand?"
- "Show me something you've written that felt 'right' -- even a Slack message or email."
Mode 3: Auto-Scrape (URL analysis)
User provides a URL. Skill uses Exa MCP to research the brand's public presence.
- Use Exa MCP to search for the brand's website content, social profiles, and public-facing copy
- Analyze homepage, about page, social bios, recent posts
- Extract voice patterns using the same DNA framework as Mode 1
- Present findings and ask user to confirm/adjust
- Generate voice profile
Mode Selection
Not sure which mode? Here's the decision tree:
- Have 3+ pieces of existing content (blog posts, emails, landing pages) → Mode 1: Extract
- Starting fresh with no content → Mode 2: Build via interview
- Have a live website URL → Mode 3: Auto-Scrape
If the brand has both a website AND existing content, prefer Mode 1 (Extract) — it captures the voice the brand already uses, which is more authentic than what a URL scan reveals.
Voice DNA Framework
Every voice maps across these axes:
Personality Spectrum
Rate each 1-10 with brief justification:
- Formal / Casual
- Reserved / Expressive
- Serious / Playful
- Traditional / Progressive
- Complex / Simple
- Cautious / Bold
Voice Markers
Sentence patterns: Average length, variation, structure (simple/compound/fragmented)
Vocabulary: Jargon level (none/light/technical), power words, forbidden words
Perspective: Person (I/we/you), distance (intimate/conversational/professional), authority (peer/mentor/expert/challenger)
The Do/Don't Table
Minimum 5 rows, maximum 10. Specific voice behaviors and example phrases.
Platform Adaptation Table
| Platform |
Tone Shift |
Length |
Special Notes |
| LinkedIn |
More professional, authority-forward |
Medium |
Lead with insight |
| Twitter/X |
Punchier, more direct |
Short |
Fragment sentences OK |
| Instagram |
Warmer, more visual language |
Short-Medium |
Casual, aspirational |
| Email |
Most personal, conversational |
Medium-Long |
First-person storytelling |
| Blog/SEO |
Most detailed, structured |
Long |
Can be more technical |
| TikTok |
Most casual, spoken-word feel |
Very short |
Conversational, raw |
| Threads |
Thoughtful, conversational |
Short-Medium |
Opinion-led |
| Bluesky |
Nuanced, substantive |
Short |
Informed, not aggressive |
Output Format
Header
BRAND VOICE PROFILE
[Brand/Person Name]
Generated [Month Day, Year]
Content
- Voice Summary (2-3 sentences capturing the essence)
- Personality Spectrum (rated axes)
- Voice Markers (sentence patterns, vocabulary, perspective)
- The Do/Don't Table
- Platform Adaptation Table
- Sample Copy (3 examples: short social post, email opener, landing page headline)
Files Saved
Write to ./brand/voice-profile.md.
What's Next
-> /positioning-angles Find your market angle
-> /direct-response-copy Write copy in your voice
-> /content-atomizer Create social content
File Output Protocol
Write ./brand/voice-profile.md with: last updated, voice summary, personality spectrum, voice markers, do/don't table, platform adaptations, sample copy.
If ./brand/voice-profile.md already exists: Read existing file, show what will change, ask for confirmation before overwriting.
Feedback
After the voice profile is saved and used in at least one downstream skill, collect feedback:
- "How did the voice feel in that [email/post/page]? (a) Nailed it, (b) Close but edited, (c) Off, (d) Haven't used yet"
- For (a) or (b): append what worked to
brand/learnings.md with date and skill name.
- For (c): ask what felt wrong, adjust the voice profile, and log the change to
brand/learnings.md.
Voice Test Loop (Required)
After generating a voice profile from ANY mode, validate before saving. Full details in references/voice-modes.md.
- Generate 3 sample paragraphs in the brand's voice: email opening, social post, landing page hero.
- Present with options: (1) Nails it → save, (2) Close but needs adjustment → ask what's off, revise, re-test, (3) Not quite right → ask for examples, rebuild relevant sections, re-test.
- After 3 rounds without confirmation, offer: save current version, one more round, or restart with a different mode.
Never skip this step. Never save a profile without user confirmation.
Anti-Patterns
| Anti-pattern |
Instead |
Why |
| Generating a generic "professional yet approachable" voice |
Extract specific patterns from real content or deep questioning |
Generic profiles are useless — every downstream skill produces the same bland output. The whole point is differentiation. |
| Skipping the Voice Test Loop |
Always validate with 3 samples before saving |
A profile that looks right on paper can sound wrong in practice. The test loop catches mismatches before they propagate to all other skills. |
| Copy-pasting platform adaptation template defaults |
Each brand has different platform behaviors — think about each one |
A casual brand might be MORE formal on LinkedIn. A formal brand might loosen up on Twitter. Defaults are wrong by definition. |
| Writing the profile without reading existing brand context |
Check positioning.md and audience.md first |
Voice should align with positioning and audience. Writing voice in isolation means rewriting it later when those files exist. |
| Overwriting an existing profile without showing diff |
Read existing file, show what will change, get confirmation |
The user may have manually tweaked the profile. Blindly overwriting loses their refinements. |
| Producing a profile with no example phrases |
On-brand and off-brand examples are required |
Examples are the fastest way for downstream skills to calibrate. Without them, other skills have to interpret abstract axes, which leads to drift. |
Invocation
This skill activates when:
- User asks "what's my voice" or "analyze my brand"
- User says copy sounds "generic" or "off-brand"
- Another skill needs voice context but ./brand/voice-profile.md doesn't exist
- User provides a URL and asks to analyze the brand
What This Skill Is NOT
- Does NOT write copy (that is /direct-response-copy)
- Does NOT define positioning (that is /positioning-angles)
- Does NOT research the audience (that is /audience-research)
- Does NOT create visual identity (that is /creative)
1---2name: brand-voice3description: Define or extract a consistent brand voice that other skills can use. Three modes: Extract (analyze existing content), Build (interview-based), Auto-Scrape (from URL). Use when copy sounds generic, when starting any new project, when voice feels inconsistent across channels, when onboarding a new brand, or when any skill needs voice-profile.md but it doesn't exist yet. This is always the first skill to run for a new project. Make sure to use this skill whenever the user mentions tone of voice, brand personality, 'my copy all sounds the same', 'how should I sound', 'analyze my website voice', 'define my tone', or anything about making content sound more human or distinctive. Even if the user just says 'my marketing sounds generic' — that's a voice problem.4---56# /brand-voice -- Brand Voice Engine78Generic copy converts worse than copy with a distinct voice. Not because the9words are different -- because the reader feels like they're hearing from a10PERSON, not a marketing team.1112This skill defines that voice. Either by extracting it from existing content,13building it strategically from scratch, or auto-scraping a URL to analyze a14brand's public presence.1516For output formatting, see /cmo [rules/output-format.md](../cmo/rules/output-format.md).1718---1920## On Activation21221. Read `brand/` context:23 - **positioning.md** — If available, use market angles to inform voice differentiation. A brand positioned as "the anti-course course" needs a different voice than one positioned as "enterprise-grade."24 - **audience.md** — If available, tailor vocabulary and tone to the target reader. Technical builders get different language than busy executives.25 - If neither exists, proceed with defaults. Ask the user about their market and audience during the question sequence.262. Check if `./brand/voice-profile.md` already exists:27 - **Exists** → Update mode: show current profile summary, ask what to change.28 - **Does not exist** → Proceed to Mode Selection below.293. Assess the user's request to determine which mode to use (Extract, Build, or Auto-Scrape).3031## Three Modes3233### Mode 1: Extract (analyze existing content)3435User provides 3-5 pieces of existing content (blog posts, emails, social posts, about page, etc).3637**The analysis process:**38391. **Read all provided content** -- look for patterns, not individual word choices402. **Identify the voice DNA:**41 - Sentence structure (short/punchy vs. flowing/complex)42 - Vocabulary level (everyday vs. technical vs. academic)43 - Emotional register (warm/friendly vs. authoritative vs. irreverent)44 - Perspective (first-person storytelling vs. objective analysis)45 - Rhythm (varies sentence length? staccato? musical?)463. **Map the personality axes** (see Voice DNA Framework below)474. **Generate the voice profile**485. **Write to ./brand/voice-profile.md**4950### Mode 2: Build (construct from scratch)5152No existing content. Build voice strategically through guided questions.5354**The question sequence:**55561. "If your brand were a person at a dinner party, how would they talk?"572. "Who is your audience? Technical builders, creative professionals, busy executives, everyday consumers, or a mix?"583. "Name 2-3 brands whose tone you admire (not their product -- their VOICE)."594. "What words should NEVER appear in your marketing? What feels off-brand?"605. "Show me something you've written that felt 'right' -- even a Slack message or email."6162### Mode 3: Auto-Scrape (URL analysis)6364User provides a URL. Skill uses Exa MCP to research the brand's public presence.65661. Use Exa MCP to search for the brand's website content, social profiles, and public-facing copy672. Analyze homepage, about page, social bios, recent posts683. Extract voice patterns using the same DNA framework as Mode 1694. Present findings and ask user to confirm/adjust705. Generate voice profile7172## Mode Selection7374Not sure which mode? Here's the decision tree:75- **Have 3+ pieces of existing content** (blog posts, emails, landing pages) → Mode 1: Extract76- **Starting fresh with no content** → Mode 2: Build via interview77- **Have a live website URL** → Mode 3: Auto-Scrape7879If the brand has both a website AND existing content, prefer Mode 1 (Extract) — it captures the voice the brand already uses, which is more authentic than what a URL scan reveals.8081---8283## Voice DNA Framework8485Every voice maps across these axes:8687### Personality Spectrum8889Rate each 1-10 with brief justification:90- Formal / Casual91- Reserved / Expressive92- Serious / Playful93- Traditional / Progressive94- Complex / Simple95- Cautious / Bold9697### Voice Markers9899**Sentence patterns:** Average length, variation, structure (simple/compound/fragmented)100101**Vocabulary:** Jargon level (none/light/technical), power words, forbidden words102103**Perspective:** Person (I/we/you), distance (intimate/conversational/professional), authority (peer/mentor/expert/challenger)104105### The Do/Don't Table106107Minimum 5 rows, maximum 10. Specific voice behaviors and example phrases.108109### Platform Adaptation Table110111| Platform | Tone Shift | Length | Special Notes |112|----------|-----------|--------|--------------|113| LinkedIn | More professional, authority-forward | Medium | Lead with insight |114| Twitter/X | Punchier, more direct | Short | Fragment sentences OK |115| Instagram | Warmer, more visual language | Short-Medium | Casual, aspirational |116| Email | Most personal, conversational | Medium-Long | First-person storytelling |117| Blog/SEO | Most detailed, structured | Long | Can be more technical |118| TikTok | Most casual, spoken-word feel | Very short | Conversational, raw |119| Threads | Thoughtful, conversational | Short-Medium | Opinion-led |120| Bluesky | Nuanced, substantive | Short | Informed, not aggressive |121122---123124## Output Format125126### Header127128```129 BRAND VOICE PROFILE130 [Brand/Person Name]131 Generated [Month Day, Year]132```133134### Content1351361. **Voice Summary** (2-3 sentences capturing the essence)1372. **Personality Spectrum** (rated axes)1383. **Voice Markers** (sentence patterns, vocabulary, perspective)1394. **The Do/Don't Table**1405. **Platform Adaptation Table**1416. **Sample Copy** (3 examples: short social post, email opener, landing page headline)142143### Files Saved144145Write to `./brand/voice-profile.md`.146147### What's Next148149```150 -> /positioning-angles Find your market angle151 -> /direct-response-copy Write copy in your voice152 -> /content-atomizer Create social content153```154155---156157## File Output Protocol158159Write ./brand/voice-profile.md with: last updated, voice summary, personality spectrum, voice markers, do/don't table, platform adaptations, sample copy.160161**If ./brand/voice-profile.md already exists:** Read existing file, show what will change, ask for confirmation before overwriting.162163---164165## Feedback166167After the voice profile is saved and used in at least one downstream skill, collect feedback:168- "How did the voice feel in that [email/post/page]? (a) Nailed it, (b) Close but edited, (c) Off, (d) Haven't used yet"169- For (a) or (b): append what worked to `brand/learnings.md` with date and skill name.170- For (c): ask what felt wrong, adjust the voice profile, and log the change to `brand/learnings.md`.171172173## Voice Test Loop (Required)174175After generating a voice profile from ANY mode, validate before saving. Full details in [references/voice-modes.md](references/voice-modes.md).1761771. Generate 3 sample paragraphs in the brand's voice: email opening, social post, landing page hero.1782. Present with options: (1) Nails it → save, (2) Close but needs adjustment → ask what's off, revise, re-test, (3) Not quite right → ask for examples, rebuild relevant sections, re-test.1793. After 3 rounds without confirmation, offer: save current version, one more round, or restart with a different mode.180181Never skip this step. Never save a profile without user confirmation.182183## Anti-Patterns184185| Anti-pattern | Instead | Why |186|-------------|---------|-----|187| Generating a generic "professional yet approachable" voice | Extract specific patterns from real content or deep questioning | Generic profiles are useless — every downstream skill produces the same bland output. The whole point is differentiation. |188| Skipping the Voice Test Loop | Always validate with 3 samples before saving | A profile that looks right on paper can sound wrong in practice. The test loop catches mismatches before they propagate to all other skills. |189| Copy-pasting platform adaptation template defaults | Each brand has different platform behaviors — think about each one | A casual brand might be MORE formal on LinkedIn. A formal brand might loosen up on Twitter. Defaults are wrong by definition. |190| Writing the profile without reading existing brand context | Check positioning.md and audience.md first | Voice should align with positioning and audience. Writing voice in isolation means rewriting it later when those files exist. |191| Overwriting an existing profile without showing diff | Read existing file, show what will change, get confirmation | The user may have manually tweaked the profile. Blindly overwriting loses their refinements. |192| Producing a profile with no example phrases | On-brand and off-brand examples are required | Examples are the fastest way for downstream skills to calibrate. Without them, other skills have to interpret abstract axes, which leads to drift. |193194## Invocation195196This skill activates when:197- User asks "what's my voice" or "analyze my brand"198- User says copy sounds "generic" or "off-brand"199- Another skill needs voice context but ./brand/voice-profile.md doesn't exist200- User provides a URL and asks to analyze the brand201202## What This Skill Is NOT203204- Does NOT write copy (that is /direct-response-copy)205- Does NOT define positioning (that is /positioning-angles)206- Does NOT research the audience (that is /audience-research)207- Does NOT create visual identity (that is /creative)
moizibnyousaf