K-Humanizer: Open-source Agent Skill for Natural Korean Writing

Description and use

K-Humanizer is an open-source Agent Skill for Cursor, Claude, and Codex that rewrites LLM-generated Korean text to sound more natural. It removes English-like phrasing, excessive formality, and abstract nouns while preserving facts, numbers, and domain terms.

Problem

Korean text generated by LLMs is often grammatically correct but sounds translated, featuring English-like word order, abstract nouns hiding actions, and excessive formality.

Outcome

The tool rewrites Korean text to sound more natural by removing abstract wording and stating actions directly, as demonstrated in the example where a formal sentence about customer inquiry classification was simplified.

Workflow

• Install the skill using npx: 'npx skills add evergreentree97/K-Humanizer --skill k-humanizer --full-depth'
• Provide the skill with the purpose of the text and its intended reader.
• Run the skill on the Korean text to revise translation-like phrasing and unnecessary formality.
• Review the output to ensure facts, numbers, names, and product terms remain unchanged.

Evidence

Before: 고객 문의 분류 체계 고도화를 통해 주간 평균 처리 시간을 18시간에서 11시간으로 크게 개선했습니다. After: 고객 문의 분류 기준을 정리해 주간 평균 처리 시간을 18시간에서 11시간으로 줄였습니다.

Source

The repository includes 200 synthetic examples across resumes, everyday writing, chat, email, documents, product copy, dialogue, and code review.

Source