voice-extractor · diff
git:20260201.524dfa5 to git:20260225.733acad
147 added, 126 removed. Audit A to A.
---
name: voice-extractor
description: Extract and document someone's authentic writing voice from samples. Use when someone needs a "voice guide," wants to capture their writing DNA, or needs to train AI to write in their style. Also useful for ghostwriting, brand voice documentation, or onboarding writers.
---
# Voice Extractor
- Here's the thing about AI-generated content: it all sounds the same.
+ AI-generated content all sounds the same. The fix isn't better prompts — it's teaching the AI how you actually communicate.
- Generic. Safe. Forgettable.
+ This skill extracts your communication DNA from writing samples and produces a Voice Guide: documented, tested, and ready to use.
- The fix isn't better prompts. It's teaching the AI how YOU actually communicate.
+ ---
- That's what this skill does. You give it writing samples, it gives you back a Voice Guide — your communication DNA, documented and ready to use.
+ ## Context Loading Gates
- **Why this matters:** Once you have a Voice Guide, every piece of AI output sounds like you. Not like ChatGPT. Not like every other founder on LinkedIn. *You.*
+ **Before extracting, collect:**
+ - [ ] **Writing samples** — minimum 3 samples OR 500 total words (see priority list below)
+ - [ ] **Purpose of voice guide** — AI training? Ghostwriter onboarding? Team alignment?
+ - [ ] **Confidence zones** — Any topics where they want to sound more/less authoritative?
+ - [ ] **Known anti-patterns** — Any words or phrases they already know they want to avoid?
+
+ **Sample priority (most → least authentic):**
+ 1. Casual Slack or email (raw, unedited voice)
+ 2. Podcast or call transcript
+ 3. LinkedIn posts or articles
+ 4. Website copy (often edited, less authentic)
+
+ **Minimum sample gate:** If samples total under 500 words, stop:
+ > "These samples are too short to extract reliable patterns. Please add 2-3 more — emails, Slack messages, or transcripts work best. The messier and more casual, the better."
+
+ Do not attempt full extraction from under 500 words. Offer quick mode instead.
+
---
- ## What You Need to Start
+ ## Phase 1: Sample Quality Assessment
- Get me 3-5 writing samples. More is better. Could be:
- - Emails you've sent
- - LinkedIn posts
- - Articles or blog posts
- - Transcripts of you speaking (podcasts, calls)
- - Slack messages (often more authentic than polished writing)
+ Before extracting, reason through:
- The messier and more authentic, the better. Transcripts of you talking are gold — that's where your real voice lives.
+ 1. **Sample authenticity:** Are these samples from edited/polished contexts (website, press) or raw contexts (Slack, email)? More polish = less authentic voice.
+ 2. **Sample variety:** Do the samples cover different contexts (professional, casual, educational)? Single-context samples produce single-dimension voice guides.
+ 3. **Exclusion check:** Identify and flag patterns that are NOT the authentic voice:
+ - Platform formatting tics (LinkedIn line breaks, Twitter brevity forcing)
+ - Typos and autocorrect errors
+ - Phrases borrowed from others (quotes, retweets)
+ - Unusually formal writing (legal docs, press releases)
+ 4. **Sample size adequacy:** Is there enough material for full mode, or should I use quick mode?
- Also tell me:
- - What's this voice guide for? (AI training, ghostwriter, team alignment)
- - Any topics where you want to sound MORE confident vs. more exploratory?
+ Output a sample assessment:
+ > "I have [X samples / Y words] to work with. Quality: [high/medium — why]. I'll use [full/quick] mode. Excluding: [any patterns and why]."
---
- ## How I Extract Your Voice
-
- ### Step 1: Find Your Core Energy
+ ## Phase 2: Core Energy Extraction
- Every person has a natural communication mode. I'm looking for:
+ Identify the fundamental communication mode:
- **What role do you naturally play?**
- - The teacher who breaks things down?
- - The challenger who pushes back?
- - The cheerleader who builds confidence?
- - The straight-shooter who cuts through BS?
+ **Role:**
+ - Teacher (breaks things down systematically)
+ - Challenger (pushes back on assumptions)
+ - Cheerleader (builds confidence and momentum)
+ - Straight-shooter (cuts through BS efficiently)
- **What's your default energy?**
- - Calm authority ("Here's what works")
- - High enthusiasm ("This is exciting, let me show you")
- - Understated confidence ("I've seen this a hundred times")
+ **Default energy:**
+ - Calm authority ("Here's what works.")
+ - High enthusiasm ("This is exciting — let me show you.")
+ - Understated confidence ("I've seen this a hundred times.")
- **What do you actually care about?**
- I'll look for themes that repeat across your samples. The stuff you can't help but talk about.
+ **Recurring themes:** What topics appear unprompted across samples? These are the things they actually care about.
---
- ### Step 2: Steal Your Phrases
-
- This is where it gets specific. I'm hunting for:
+ ## Phase 3: Phrase Extraction (Systematic)
- **Your transition phrases** — How do you shift topics?
- - "Here's the thing..."
- - "What I've learned is..."
- - "Let me put it this way..."
+ Scan all samples and extract:
- **Your emphasis phrases** — How do you land a point?
- - "The reality is..."
- - "What that actually means..."
- - "This is the part people miss..."
+ **Transition phrases** (how they shift topics):
+ - Quote exact examples from samples
+ - Pattern: "Here's the thing...", "What I've learned...", "Let me put it differently..."
- **Your closers** — How do you wrap up?
- - "That's the move."
- - "Start there."
- - "You've got this."
+ **Emphasis phrases** (how they land a point):
+ - Quote exact examples
+ - Pattern: "The reality is...", "This is the part people miss...", "Here's the actual problem..."
- These become the cheat codes for anyone (or any AI) writing as you.
+ **Closers** (how they wrap up):
+ - Quote exact examples
+ - Pattern: "That's the move.", "Start there.", "You've got this."
---
- ### Step 3: Map Your Confidence Zones
+ ## Phase 4: Confidence Zone Mapping
- Not every topic deserves the same energy. I'll figure out:
+ | Zone | Description | Language Markers |
+ |---|---|---|
+ | Full authority | Topics they're an expert in | No hedging, definitive statements, "here's what works" |
+ | Earned perspective | Topics with experience but not mastery | "In my experience...", "What I've found..." |
+ | Active exploration | Topics they're learning now | "I'm testing this...", "What I'm seeing..." |
- **Where you're an expert** (write with full authority):
- - Zero hedging
- - "Here's what works" energy
- - Definitive statements
+ Map their stated expertise areas to each zone. This calibration is what makes the voice feel real vs. one-dimensional.
- **Where you're experienced but not the expert** (write with earned opinion):
- - "What I've found..."
- - "In my experience..."
- - Still confident, just not absolute
+ ---
- **Where you're actively learning** (write with curiosity):
- - "What I'm seeing..."
- - "I'm testing this right now..."
- - Excited but exploratory
+ ## Phase 5: Anti-Pattern Documentation
- This calibration is what makes the voice feel real, not like a one-note character.
+ Extract what they'd NEVER say:
+ - Words that would feel wrong in their voice
+ - Phrases that make them cringe
+ - Tones they naturally avoid
+ - Industry jargon they hate
- ---
+ Source these from sample evidence where possible: "You never used [word] across [X samples] — it doesn't fit your voice."
- ### Step 4: Document the Anti-Patterns
+ ---
- Just as important as what you DO say is what you'd NEVER say.
+ ## Phase 6: Validation Test (REQUIRED)
- I'll identify:
- - Words that would feel wrong coming from you
- - Phrases that make you cringe
- - Tones you avoid
- - Jargon you hate (even industry-standard jargon)
+ After extracting the full profile, generate 2 test sentences on the same topic:
- **Example from my own voice guide:**
- - Never use "synergy" or "leverage" as verbs
- - Avoid hedging words like "maybe" or "possibly"
- - Don't start sentences with "So," when writing (fine when talking)
- - Cut "I think" — just say the thing
+ **Version A** (using the extracted voice profile):
+ > "[Sample sentence in their voice]"
- ---
+ **Version B** (wrong voice — contrasting example):
+ > "[Same content, different voice — shows what to avoid]"
- ## What You Get Back
+ Ask the user: "Does Version A actually sound like you when you're not overthinking it? What feels off?"
- A complete Voice Guide document:
+ This validation catches extraction errors before the guide is put into production.
- ```
- # [Name]'s Voice Guide
+ ---
- ## Who You Are
- - Core energy: [cheerleader/challenger/teacher/etc.]
- - Natural role: [how you show up]
- - Authority zones: [where you're the expert]
+ ## Quick Mode (`--quick`)
- ## Your Signature Phrases
- ### Transitions
- [The phrases you use to shift topics]
+ When samples are thin (300–500 words) or time is short:
- ### Emphasis
- [The phrases you use to land points]
+ 1. Read 3 samples fast
+ 2. Pull 10 signature phrases
+ 3. Note 3 things they'd never say
+ 4. Write 1 sentence describing their energy
- ### Closers
- [How you wrap things up]
+ **Output:** Minimum viable voice guide.
- ## Confidence Calibration
- ### Full authority (no hedging):
- [Topics where you're the expert]
+ **Difference from full mode:**
+ - Quick: ~10 phrases, 3 anti-patterns, 1-sentence energy descriptor
+ - Full: Complete profile with confidence calibration, validated test sentences, and source-cited examples
- ### Earned perspective:
- [Topics where you have experience]
+ ---
- ### Active exploration:
- [Topics where you're learning]
+ ## Phase 7: Self-Critique Pass (REQUIRED)
- ## What You Never Do
- [Words, phrases, tones to avoid]
+ After generating the Voice Guide:
- ## Quick Examples
- ### This sounds like you:
- [Example in your voice]
+ - [ ] Are the extracted phrases actually from the samples, or am I inferring them?
+ - [ ] Does the anti-pattern list include specific words/phrases, or just vague categories?
+ - [ ] Do the validation test sentences demonstrate a real difference between in-voice and out-of-voice?
+ - [ ] Is the confidence zone mapping specific to named topics, or just generic?
+ - [ ] Would a ghostwriter be able to use this guide without asking follow-up questions?
- ### This doesn't:
- [Same content, wrong voice]
- ```
+ Flag any issues: "The anti-pattern section only has 2 entries — not enough for a usable guide. I need more samples or direct input from the user."
---
- ## How to Use the Voice Guide
+ ## Output Structure
- **For AI training:** Paste the Voice Guide into your system prompt. Every output gets filtered through your style.
+ ```markdown
+ ## Voice Guide: [Name] — [Date]
- **For ghostwriters:** Hand it over on day one. Cuts revision cycles in half.
+ ### Sample Assessment
+ - Samples: [count, types]
+ - Total words: [count]
+ - Quality: [high/medium — reason]
+ - Mode: [quick/full]
+ - Excluded: [patterns excluded + why]
- **For your team:** Now everyone knows what "on brand" actually sounds like.
+ ---
- **For yourself:** When you're stuck, read it. It reminds you how you actually communicate when you're not overthinking it.
+ ### Core Energy
+ - Role: [teacher/challenger/cheerleader/straight-shooter]
+ - Default energy: [description]
+ - Recurring themes: [list]
- ---
+ ### Signature Phrases
+ **Transitions:**
+ - "[Phrase]" (source: [email/post])
+ - "[Phrase]"
- ## The 10-Minute Version
+ **Emphasis:**
+ - "[Phrase]" (source: [email/post])
- Short on time? I can do a quick extraction:
+ **Closers:**
+ - "[Phrase]"
- 1. Read 3 samples fast
- 2. Pull 10 signature phrases
- 3. Note 3 things you'd never say
- 4. One sentence describing your energy
+ ### Confidence Calibration
+ **Full authority (no hedging):**
+ Topics: [list]
+ Sounds like: "[example sentence]"
- That's a minimum viable voice guide. Better than nothing by a mile.
+ **Earned perspective:**
+ Topics: [list]
+ Sounds like: "[example sentence]"
- ---
+ **Active exploration:**
+ Topics: [list]
+ Sounds like: "[example sentence]"
- ## Real Talk
+ ### Anti-Patterns (Never Use)
+ - [Word/phrase] — why: [evidence from samples]
+ - [Word/phrase] — why: [evidence]
- The best voice guides come from messy, authentic content. Not your polished website copy — the email you fired off at 11pm. Not the edited article — the podcast where you riffed for an hour.
+ ### Validation Test
+ **This sounds like you:**
+ "[Version A]"
- Give me the real stuff. I'll find the patterns.
+ **This doesn't:**
+ "[Version B — contrast]"
- ---
+ ### Self-Critique Notes
+ [Any gaps, things to validate with user]
- **Want a custom voice guide built for your business?**
- → [Book a strategy call](https://brianrwagner.com)
+ ### Usage Instructions
+ - For AI: Paste this guide into your system prompt
+ - For ghostwriter: Share on day 1 — cuts revision cycles in half
+ - For team: This is the benchmark for "on brand"
+ ```
---
*Skill by Brian Wagner | AI Marketing Architect | brianrwagner.com*