Natural writing humanization

Use when editing Korean or English prose so it reads naturally while preserving facts, meaning, technical terms, and author intent.From its SKILL.md

Install

npx -y skills add jukrap/ai-agent-playbook --skill natural-writing-humanization

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

1.5 KB, 284 tokens by cl100k_base, as published. Nobody here has run it

Improve prose quality without changing what the document means.

Identify language, audience, source type, and whether this is a light edit, translation cleanup, or full rewrite.
Preserve facts, numbers, names, dates, warnings, release scope, legal meaning, and technical contracts before changing style.
Run aapb writing naturalness-check <target> --path <file> --lang auto --engine auto --json when a local file is available.
Remove machine-like phrasing, translationese, forced symmetry, inflated claims, and repetitive rhythm.
Keep the writer's voice, domain vocabulary, and useful bluntness; do not over-polish into generic marketing copy.
Refuse requests framed as detector bypass, evaluation evasion, fake authorship, or policy avoidance.

Read references/korean-naturalness-patterns.md for Korean translationese, English-term density, and natural Korean rewrite checks.

Read references/english-naturalness-patterns.md for common AI-writing phrases, padded rhythm, and plain English rewrite checks.

Read references/voice-fidelity-and-boundaries.md for meaning preservation, safety boundaries, and over-edit prevention.

Read references/review-rubric.md before finalizing docs, README text, PR bodies, release notes, or translations.

7.9 KB alongside SKILL.md