---

name: voice-learning description: Knowledge base for analyzing and replicating writing voice. Use when learning voice patterns or generating voice-matched content.

Voice Learning Skill

This skill provides frameworks and techniques for analyzing writing voice and ensuring content generation matches a learned voice profile.

Voice Analysis Framework

The 5 Dimensions of Voice

1. Lexical Voice - Word choice

   • Vocabulary complexity (grade level)
   • Industry jargon usage
   • Power words and phrases
   • Words consciously avoided

2. Syntactic Voice - Sentence structure

   • Average sentence length
   • Sentence variety (simple, compound, complex)
   • Use of fragments for effect
   • Paragraph length preferences

3. Rhetorical Voice - Persuasion patterns

   • Primary appeal (ethos, pathos, logos)
   • Use of questions
   • Analogy and metaphor frequency
   • Storytelling vs. data-driven

4. Tonal Voice - Emotional coloring

   • Formality spectrum (1-10)
   • Warmth/distance
   • Confidence level
   • Humor integration

5. Structural Voice - Organization

   • Opening patterns (hook types)
   • Transition preferences
   • Closing/CTA style
   • Use of formatting elements

Voice Extraction Techniques

From Transcripts

Transcripts reveal natural, unfiltered voice:

• Listen for filler phrases - These often carry personality
• Note explanation patterns - How do they break down complex ideas?
• Capture spontaneous analogies - Unplanned comparisons reveal thinking
• Track energy shifts - What topics generate enthusiasm?

From Written Content

Published writing shows intentional voice:

• Analyze hooks - First sentences reveal attention strategy
• Study transitions - How do ideas connect?
• Examine conclusions - What's the signature close?
• Note formatting - Headers, bullets, bold usage

From Social Media

Social content shows engagement voice:

• Opening hooks - How is scrolling stopped?
• Thread structure - How are ideas chunked?
• Engagement prompts - Questions, CTAs
• Community signals - In-group references

Voice Matching Checklist

Before generating content, verify:

• Vocabulary matches profile (no words from "avoid" list)
• Sentence rhythm matches (length, variety)
• Opening style matches learned pattern
• Tone is consistent (formality, warmth)
• Frameworks/analogies are in-character
• CTA style matches preference

Memory Storage Schema

Store voice patterns with these categories:

CopyCategories: ["voice-profile", "<dimension>"]

Dimensions:
- "lexical" - Word choice patterns
- "syntactic" - Sentence structure
- "rhetorical" - Persuasion patterns
- "tonal" - Emotional coloring
- "structural" - Organization patterns
- "topic-specific" - Voice variations by subject
- "platform-specific" - Voice variations by channel

Voice Injection for Content Generation

When generating content in user's voice:

1. Load Profile

   CopyRecall voice-profile patterns from memory
   

2. Apply Constraints

   • Use vocabulary from profile
   • Match sentence structure patterns
   • Apply tonal settings
   • Follow structural preferences

3. Verify Match

   • Read output aloud mentally
   • Check against "avoid" list
   • Confirm opening/closing match patterns

4. Note Deviations

   • If topic requires different voice, note it
   • Flag for user review if uncertain

Platform-Specific Voice Adjustments

LinkedIn

• Slightly more formal
• Professional but personable
• Thought leadership framing
• Engagement-focused endings

Twitter/X

• More casual, punchy
• Thread-optimized structure
• Hook-heavy openings
• Community engagement signals

Blog/Long-form

• Full voice expression
• Story-driven when appropriate
• Technical depth as needed
• Signature frameworks

Email

• Direct and action-oriented
• Relationship-appropriate formality
• Clear next steps

Common Voice Pitfalls

1. Over-formalizing - AI tendency to sound corporate
2. Losing quirks - Removing personality for "polish"
3. Inconsistent tone - Shifting formality mid-piece
4. Wrong jargon - Using terms outside their vocabulary
5. Generic CTAs - Not matching their engagement style

Voice Evolution

Voice changes over time. Periodically:

• Analyze recent content (last 90 days)
• Compare to older profile entries
• Update patterns that have shifted
• Archive outdated patterns (don't delete)