This skill transforms machine-like text into natural human prose. It identifies common AI-writing patterns (overstatement, promotional language, repetitive three-item lists, overuse of dashes and filler phrases) and rewrites problematic passages while preserving meaning and tone. The skill includes rules, examples, and a checklist to guide automated or human-in-the-loop editing.
Apply this skill when you need to neutralize AI artifacts from drafts, prepare content for publication, or improve readability and voice. Useful for editors, content teams, and moderation workflows where writing must be localized to a human style or matched to a target tone.
Works with general-purpose text-editing agents (Claude, GPT-style assistants, transformer-based editors) that support read/write/edit actions.
A Chinese-language skill that provides a comprehensive guide for identifying and removing AI-generated writing patterns from text. It covers 24 distinct AI writing patterns (exaggerated symbolism, promotional language, vague attribution, rule-of-three overuse, etc.) with before/after examples, a processing workflow, and a quality scoring rubric. No scripts included — purely prompt-based, using only Read/Write/Edit/AskUserQuestion tools. Well-structured and detailed, but entirely in Chinese which significantly limits its audience.
Clean skill with no security concerns. The SKILL.md body was null in the DB fetch, had to retrieve from GitHub raw URL. The content is high-quality instructional material for a specific niche (Chinese-language AI text cleanup). Architecture could be improved by splitting the long pattern catalog into a references/ file to keep SKILL.md leaner.
