| name | create-writing-dna |
| description | Use when someone wants to capture their personal writing voice/style so Claude can write as them — building a "writing DNA" from their past emails, chat, posts, and docs across multiple media, and generating a portable per-person voice skill. Also use to refresh or upgrade an existing voice profile. |
Create Writing DNA
Overview
Extract a person's writing voice from their real samples across multiple media, then emit a portable per-person voice skill — a <name>-voice/ folder containing SKILL.md (trigger), writing-dna.md (the voice profile), and generating.md (the generation workflow). The folder has zero external dependencies: copy it into any skills/ directory and the voice works standalone.
When to use
- "Help Claude write like me" / "in my voice" / "match my writing style"
- Building or refreshing a personal/teammate voice profile from past writing
- Bilingual writers who need code-switching captured (e.g. PT/EN)
Not for: one-off ghostwriting with no profile, or generic marketing copy.
Workflow
- Assemble the corpus — discover the user's own writing generically. See
references/source-discovery.md.
- Run the passes — Pass 0-3 (or quick mode). See
references/extraction-process.md.
- Apply the 8 dimensions + VOICE/PLATFORM/BORDERLINE classification. See
references/analysis-dimensions.md.
- Personalize the ban-list from the base list. See
references/llm-isms.md.
- Emit the voice skill from
templates/voice-skill/ (fill SKILL.md.tmpl, write writing-dna.md, copy generating.md verbatim). See the "Emit" section of references/extraction-process.md.
Key rules
- Ban-lists first in the DNA file; earlier constraints bind hardest.
- Cite evidence for every pattern ("seen in N/M ").
- Only the user's own, human-written content — never received text, never AI-assisted samples.
- Never hard-code tool/skill/MCP names — discover what's available at runtime.
- One DNA file for all languages, with language-specific subsections.
Output
A portable <name>-voice/ skill folder (default ~/.claude/skills/<name>-voice/). This is personal output — it is not committed to this repo.
