Humanize

Entry command that runs the humanize-korean pipeline to polish Korean AI-written text into natural prose.

What it does

This skill is an entry command that takes Korean text or a .txt/.md file path and invokes the humanize-korean pipeline to rewrite it so it reads less like AI-generated writing. It loads file contents when a path is given, follows the pipeline's phases, and chooses a light, standard, or heavy route via a route hint, forcing the heavy route when --strict is passed. It returns a one-line status with change rate, grade and self-check result, the rewritten text, before/after detection counts by category, and three to five change highlights. It also supports genre, intensity and minimum-severity options and suggests a second pass when the grade is B or lower.

When to use it

Use it when you have Korean text that sounds machine-generated and want it rewritten into more natural prose. It fits requests to remove AI writing tells from Korean drafts, columns, reports, blog posts or formal text. It is triggered by the /humanize command.

Requirements

Requires the humanize-korean skill and its reference files (ai-tell-taxonomy.md, rewriting-playbook.md) to be available. It ships no scripts and is instructions only; file input is limited to .txt and .md paths.