natural-chinese-copy-core

Use this whenever the user uses the stable Chinese trigger phrase 中文才是最屌的， followed by a request, for example 中文才是最屌的，帮我写个稿子. Also use whenever the task involves 中文文案, 中文写作, 中文稿, 稿子, 稿件, 写文案, 写稿子, 写几篇, 写一版, 改稿, 改文案, 润色文案, 审稿, 评稿, 即评, 推文, 帖子, 社媒文案, 中文介绍, or requests like 根据资料/链接/项目/README/GitHub repo/素材 帮我写个文案/稿子/介绍/推荐/摘要. Also use when Chinese copy must avoid AI味, 翻译腔, 英文语感, 模板感, fake conversational tone, or English-shaped sentence logic, or when users rate copy, 标出坏句, 做坏句反馈, mark bad sentences, compare draft quality, or want feedback saved as samples and rule-evolution proposals. It provides Chinese-first fact restatement, sentence-structure critique, register selection, colloquial sentence shaping, anti-translationese checks, batch repetition review, feedback capture, and research-backed rule distillation. Other skills should call this core instead of copying long Chinese writing rules. It does not gather sources, generate images, publish content, own product-specific facts, or automatically turn ratings in