短剧剧本与网文小说专属活人感打磨 Skill，用于清理 AI 味、模板腔、假精确、万能安抚腔和过度升华。
中英文学术写作去AI化与去防御性表述 Agent Skill · De-AI & de-hedge academic prose (zh/en) for Claude Code, Codex, pi and other Agent Skills agents
Easywrite - make AI-assisted academic writing read human while staying evidence-true, in EN + 中文. Claim-evidence discipline, detector science, bilingual de-slop, integrity firewall.
deai — de-AI writing for Chinese text. One self-contained Agent Skill: strip AI flavor from web novels, fiction, articles, titles. Clean/Refactor/Detect/Write modes; layered 5-step pipeline; graded A/B/C rules; merges human-writing, lieflat, sepia, oh-story.
De-AI Writing · 中文写作去 AI 味总控｜中文写作去 AI 味总控。两大功能：①修改已有文章——先定文体，再把 AI 生成感量化为客观指标（词表密度、句式重复率、模板腔密度、结构指纹），逐处核对后给改写方案，附完整性守卫；②从零生成文案——先逐步与用户确认文体、文风、受众、篇幅等全部要素，严禁擅自推测补料，确认完才动笔。判定标准分层：主矛盾（结论先行、防御性写作、伪排比、细节不服务上下文）优先于次要矛盾（词表句式）。覆盖七大类中文文体（实用说明、专业论证、商务事务、营销转化、叙事创作、社交表达、法规制度）四十余个子类。触发词：去 AI 味、降 AI 率、AI 感、太像 AI 写的、人机味、不像人话、润色、改写、humanize、论文降AI率、公众号文案、帮
Chinese-first humanized writing Skill for ChatGPT, Codex, Claude Code and Cursor
Linter and Claude skill for the text between you and a model. Checks prompts for the ambiguity a model will guess through, and model output for invisible watermark characters, machine rhythm, and AI tells. English and Chinese.
My personal collection of portable SKILL.md agent skills and tooling for Claude Code, Cursor & Codex — currently focused on Traditional Chinese (Taiwan) writing, de-AI editing, and business docs.
Chinese text de-AI-ifier: strips the AI flavor from Chinese writing using stylistics, lexicology and rhetoric research instead of banned-word lists. 中文去AI味改写器：九指标体检 + 定点干预 + 语义保真
基于本地 TTS 发声与生理呼吸节奏的去 AI 味双子 Agent 工作流：因声求气，盲听重构。
To associate your repository with the de-ai topic, visit your repo's landing page and select "manage topics."