Distill a Writer's Voice: Writing DNA Skill for Agents

Where this fits

You need a new article for a Chinese author's newsletter in their exact voice. Reading 20+ posts manually, then summarizing quirks, is slow and inconsistent. This skill turns that corpus into a reusable Writing-DNA.md rule set an agent can load every time it writes.

Why agents benefit

• It ingests at least 20 complete articles and extracts concrete rules for language, structure, topic logic, source strategy, cognitive frames, and visual style — not vague "tone words".
• It separates conversation language from output language, so a Chinese-speaking user can get English deliverables with proper filenames.
• It reads the distilled DNA plus 5 relevant source articles before composing, giving the agent grounded stylistic memory instead of one-shot prompting.
• It supports comparative analysis across authors or accounts, enabling competitive editorial audits.
• Output is structured Markdown templates, so CI pipelines or multi-agent workflows can consume it as a stable artifact.

Practical scenarios

Brand account voice migration

A brand moving from an agency to an internal content team needs to preserve its social/column voice. Run the skill on the account's 20 best-performing posts, commit Writing-DNA.md to the repo, and every future draft starts from the same editorial constraints.

Ghostwriting for a solo author

An author with a personal blog wants an agent to draft newsletter issues. Distill their archive, then use the DNA file as the system prompt reference for topic selection, sentence rhythm, and how they structure arguments.

Editorial benchmark analysis

A publication comparing itself to a rival can run this skill on both archives. The resulting structure and cognitive-framework comparisons reveal patterns the rival is using to drive engagement.

Add it to your agent workflow

Install with:

npx skills add larashero3-dotcom/writing-dna-skill --skill writing-dna-skill

Then provide a directory of at least 20 full articles and ask:

Distill the writing DNA from ./articles and generate Writing-DNA.md.

The agent will produce the language, structure, cognitive-framework, and visual-style files, then integrate them into one actionable rule set.

Compare before adopting

Before making this your default style-analysis skill, compare:

• Quality signals: Does it actually generate non-generic rules from real articles? Run it on a corpus you know well.
• Maintenance freshness: Check the GitHub repo for recent commits and issue responses.
• Alternatives: Other skills may summarize tone but not produce structured, article-grounded DNA files.
• Workflow fit: Does the artifact format slot into your existing content pipeline and language conventions?

Why it is worth tracking

With 1,410 stars and a clear, repeatable methodology, writing-dna-skill is gaining traction among content-agent builders who need consistency at scale. Track it if your agents write for human authors or brands; evaluate it again when the repo adds more templates or multilingual corpus support. For a catalog of similar agent skills, see https://www.openagentskill.com/skills/larashero3-dotcom-writing-dna-skill.