Ai Proofreading
- 67 installs
- 58 repo stars
- Updated April 20, 2026
- alchaincyf/glm-claude
ai-proofreading is a Claude Code skill that runs a three-pass review over text to remove AI-writing patterns and lower its AI-detection score below 30%.
About
A Claude Code skill that runs a three-pass proofreading process over drafted text to strip out mechanical AI-writing patterns and add a human voice. A developer or writer uses it after a draft is written to make content read more naturally and lower its AI-detection score. It works on Chinese-language content and enforces a fixed set of structure, style, and detail checks.
- Three-pass proofreading loop (content, style, details) to remove AI-writing tells
- Targets lowering AI-detection score from 80-100% down to under 30%
- Chinese-language rewrite checklist for more natural, human-sounding copy
Ai Proofreading by the numbers
- 67 all-time installs (skills.sh)
- Ranked #519 of 853 Sales & Marketing skills by installs in the Skillselion catalog
- Data as of Jul 28, 2026 (Skillselion catalog sync)
ai-proofreading capabilities & compatibility
- Capabilities
- ai taste proofreading · video script collaborial · article to x
- Use cases
- copywriting
What ai-proofreading says it does
系统化降低AI检测率至30%以下，通过三遍审校流程（内容、风格、细节）增加人味。
通过系统化的三遍审校流程，将内容的AI检测率从80-100%降低至30%以下，让文章更具人味和真实性。
npx skills add https://github.com/alchaincyf/glm-claude --skill ai-proofreadingAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 67 |
|---|---|
| repo stars | ★ 58 |
| Last updated | April 20, 2026 |
| Repository | alchaincyf/glm-claude ↗ |
What it does
Rewrite AI-drafted Chinese content through a three-pass review to reduce AI-detection score and make it read like a human wrote it.
Who is it for?
Writers polishing AI-drafted Chinese articles before publishing.
When should I use this skill?
When content feels too much like AI wrote it or needs a more human, colloquial tone.
What you get
Content reads more naturally and its AI-detection score drops below 30% while keeping the original meaning.
- rewritten humanized article text
By the numbers
- Three-pass review (content, style, details)
- Target AI-detection rate under 30%
- At least 3 personalized expressions added
Files
AI味审校
何时使用此Skill
当用户出现以下任一需求时，使用此技能：
- 提到内容"AI味太重"、"像AI写的"、"太机械"
- 要求"降低AI检测率"、"更像人写的"、"自然一些"
- 需要"口语化"、"有温度"、"接地气"的表达
- 内容被AI检测工具标记为高AI概率
- 希望文章更具个性和真实感
核心目标
通过系统化的三遍审校流程，将内容的AI检测率从80-100%降低至30%以下，让文章更具人味和真实性。
执行流程
第一遍：内容结构调整 (P0级检查)
目标：修复明显的AI结构特征
检查要点： 1. 避免教科书式结构
- ❌ "首先...其次...最后..."的机械结构
- ✅ 使用自然的转折和过渡
2. 避免定义式开头
- ❌ "什么是XX？XX是指..."
- ✅ 直接切入主题或使用引子开头
3. 避免总结式结尾
- ❌ "综上所述，总而言之..."
- ✅ 自然收尾或留有余韵
4. 增加个人化表达
- 插入个人观点、经历、感受
- 使用"我觉得"、"在我看来"等表达
第二遍：语言风格优化 (P1级检查)
目标：让语言更自然、更有温度
检查要点： 1. 替换书面化词汇
- ❌ "进行"、"实现"、"提供"
- ✅ "做"、"完成"、"给"
2. 增加口语化表达
- ❌ "该产品具备..."、"此方法..."
- ✅ "这个产品能..."、"这个方法..."
3. 使用生活化比喻
- 结合日常生活经验打比方
- 避免过于专业的抽象表达
4. 调整句式节奏
- 长短句结合，避免全是长句
- 适当使用设问、感叹等句式
第三遍：细节润色 (P2级检查)
目标：增加真实感和可信度
检查要点： 1. 增加具体细节
- 补充时间、地点、人物等具体信息
- 使用数字、百分比等量化表达
2. 添加情绪色彩
- 适当表达惊讶、失望、兴奋等情绪
- 使用感叹词、语气词增加感染力
3. 优化连接词
- 减少"因此"、"然而"、"另外"等正式连接词
- 使用"不过"、"而且"、"对了"等口语连接
4. 检查逻辑合理性
- 确保论证过程自然流畅
- 避免过于完美的逻辑闭环
质量标准
- AI检测率 < 30%
- 保持原意不变
- 增加至少3个个人化表达
- 语言风格与目标读者匹配
--- 最后更新：2024年
Related skills
Sales EnablementGenerate high-quality sales collateral, pitch decks, one-pagers, objection handlers, demo scripts and playbooks that help close deals.85.9k43kChurn PreventionDesign cancel flows, dunning, and health-score tactics to cut voluntary and failed-payment churn on a subscription product.85.5k43kRevopsDesign and optimize revenue operations, lead lifecycle management, and marketing-to-sales handoff processes.82.4k43k
Cross Border EcommerceEvaluate and prioritize international markets before committing engineering and marketing resources to global expansion.62k581Community MarketingGet structured, battle-tested guidance when choosing community platforms, diagnosing engagement problems, or scaling an early-stage user community.54.2k43kCo MarketingSystematically identify and brainstorm campaigns with complementary products that share the same audience but solve different problems.45.4k43kFAQ
What AI-detection rate does this skill target?
It aims to bring content from 80-100% AI-detection down to under 30% while keeping the original meaning.
How does the rewrite work?
It runs three passes: content structure (P0), language style (P1), and detail polish (P2), each with specific checks.