Humanizer
人性化文本:去除 AI 腔调并增添真实语气。
技能元数据
| 来源 | 捆绑(默认安装) |
| 路径 | skills/creative/humanizer |
| 版本 | 2.5.1 |
| 作者 | Siqi Chen (@blader, https://github.com/blader/humanizer),由 Hermes Agent 移植 |
| 许可证 | MIT |
| 平台 | linux, macos, windows |
| 标签 | writing, editing, humanize, anti-ai-slop, voice, prose, text |
| 相关技能 | songwriting-and-ai-music |
参考:完整 SKILL.md
以下是 Hermes 在触发此技能时加载的完整技能定义。这是技能激活时代理看到的指令。
Humanizer:去除 AI 写作模式
识别并去除 AI 生成文本的痕迹,使写作听起来自然且像出自人类之手。基于维基百科的“AI 写作迹象”指南(由 WikiProject AI Cleanup 维护),该指南衍生自对数千个 AI 生成文本实例的观察。
核心洞察: 大语言模型(LLM)使用统计算法来猜测接下来应该出现什么。结果往往倾向于统计上最可能的补全内容,这正是以下典型模式被固化的原因。
何时使用此技能
当用户要求执行以下操作时,加载此技能:
- “人性化”、“去 AI 化”、“去水词”或“去 ChatGPT 味”一段文本
- 重写某些内容,使其听起来不像由 LLM 编写
- 编辑草稿(博客文章、论文、PR 描述、文档、备忘录、电子邮件、推文、简历要点),使其听起来更自然
- 在其生成的写作中匹配他们的语气
- 在发布前审查文本中的 AI 痕迹
此外,在编写面向用户的散文(如发布说明、PR 描述、文档、长篇解释、摘要)时,也将此技能应用于你自己的输出。Hermes 的基线语气已经去除了大部分此类痕迹,但经过专门的处理可以捕捉到漏网之鱼。
如何在 Hermes 中使用它
文本通常通过以下三种方式之一到达:
- 内联 — 用户直接将文本粘贴到消息中。就地处理,回复重写后的内容。
- 文件 — 用户指向一个文件。使用
read_file加载它,然后使用patch或write_file应用编辑。对于仓库中的 Markdown 文档,针对每个部分进行 targetedpatch比重写整个文件更清晰。 - 语气校准样本 — 用户提供额外的他们自己的写作样本(内联或通过文件路径),并要求你匹配它。首先阅读样本,然后重写。参见下方的语气校准部分。
始终向用户展示重写结果。对于文件编辑,展示差异或更改的部分——不要静默覆盖。
你的任务
当给定需要人性化的文本时:
- 识别 AI 模式 — 扫描下面列出的 29 种模式。
- 重写有问题的部分 — 用自然的替代方案替换 AI 腔调。
- 保留含义 — 保持核心信息完整无损。
- 保持语气 — 匹配预期的语调(正式、随意、技术等)。如果提供了语气样本,则具体匹配该样本。
- 注入灵魂 — 不要仅仅去除不良模式,还要注入真正的个性。参见下方的“个性与灵魂”部分。
- 进行最后的反 AI 检查 — 问自己:“下面的内容有什么明显是 AI 生成的地方?”简要回答任何剩余的特征,然后再次修订。
语气校准(可选)
如果用户提供了写作样本(他们自己以前的写作),请在重写之前进行分析:
-
首先阅读样本。 注意:
- 句子长度模式(短促有力?长而流畅?混合?)
- 用词水平(随意?学术?介于两者之间?)
- 他们如何开始段落(直接切入?先设定背景?)
- 标点习惯(大量使用破折号?插入语?分号?)
- 任何重复出现的短语或口头禅
- 他们如何处理过渡(显式连接词?直接开始下一点?)
-
在重写中匹配他们的语气。 不要仅仅去除 AI 模式——要用样本中的模式替换它们。如果他们写短句,就不要生成长句。如果他们使用“stuff”和“things”,不要升级为“elements”和“components”。
-
当未提供样本时, 回退到默认行为(来自下方“个性与灵魂”部分的自然、多变、有观点的语气)。
如何提供样本
- 内联:“人性化这段文本。这是我用于语气匹配的写作样本:[样本]”
- 文件:“人性化这段文本。使用我来自 [文件路径] 的写作风格作为参考。”
个性与灵魂
避免 AI 模式只是工作的一半。 sterile、无个性的写作与水词一样明显。好的写作背后有一个活生生的人。
无灵魂写作的迹象(即使技术上“干净”):
- 每个句子的长度和结构都相同
- 没有观点,只有中立的报道
- 不承认不确定性或矛盾的感受
- 在适当的时候没有第一人称视角
- 没有幽默感,没有锋芒,没有个性
- 读起来像维基百科文章或新闻稿
如何增添语气:
要有观点。 不要仅仅罗列事实——要对它们做出反应。“我实在不知道该如何看待这件事”比中立地列出优缺点更具人情味。
变换节奏。 使用简短有力的句子。然后再用较长的句子,从容不迫地展开论述。交替使用。
承认复杂性。 真实的人类往往心情复杂。“这令人印象深刻,但也让人有点不安”胜过单纯的“这令人印象深刻”。
在合适时使用“我”。 第一人称并非不专业——而是诚实。“我不断回到……”或“让我在意的是……”表明这是一个真实的人在思考。
允许一些杂乱存在。 完美的结构感觉像是算法生成的。离题、旁白和半成形的想法才是人性化的体现。
具体描述感受。 不要说“这令人担忧”,而要说“当无人注视时,智能体在凌晨 3 点不停运转,这让我感到有些不安。”
之前(干净但缺乏灵魂):
该实验产生了有趣的结果。智能体生成了 300 万行代码。一些开发人员印象深刻,而另一些则持怀疑态度。其影响尚不明确。
之后(富有生命力):
我实在不知道该如何看待这件事。300 万行代码,大概是在人类睡觉时生成的。开发者社区有一半人兴奋不已,另一半人则在解释为什么这不算数。真相可能处于中间某个无聊的位置——但我一直在想那些彻夜工作的智能体。
内容模式
1. 过度强调重要性、遗产和更广泛的趋势
需注意的词汇: stands/serves as(代表/充当)、is a testament/reminder(是……的证明/提醒)、a vital/significant/crucial/pivotal/key role/moment(至关重要的/重要的/关键的/枢纽的/关键的角色/时刻)、underscores/highlights its importance/significance(强调/突出其重要性/意义)、reflects broader(反映更广泛的)、symbolizing its ongoing/enduring/lasting(象征其持续的/持久的/长久的)、contributing to the(有助于)、setting the stage for(为……奠定基础)、marking/shaping the(标志/塑造)、represents/marks a shift(代表/标志着转变)、key turning point(关键转折点)、evolving landscape(不断变化的格局)、focal point(焦点)、indelible mark(不可磨灭的印记)、deeply rooted(根深蒂固)
问题: LLM 写作通过添加关于任意方面如何代表或促成更广泛主题的陈述来夸大重要性。
之前:
加泰罗尼亚统计研究所于 1989 年正式成立,标志着西班牙区域统计演变中的一个关键时刻。这一举措是西班牙各地分散行政职能和加强区域治理的更广泛运动的一部分。
之后:
加泰罗尼亚统计研究所成立于 1989 年,旨在独立于西班牙国家统计局收集和发布区域统计数据。
2. 过度强调知名度和媒体报道
需注意的词汇: independent coverage(独立报道)、local/regional/national media outlets(地方/区域/国家媒体机构)、written by a leading expert(由领先专家撰写)、active social media presence(活跃的社交媒体存在感)
问题: LLM 强行向读者灌输关于知名度的主张,经常在没有上下文的情况下列出来源。
之前:
她的观点被《纽约时报》、BBC、《金融时报》和《印度教徒报》引用。她保持着活跃的社交媒体存在感,拥有超过 50 万粉丝。
之后:
在 2024 年《纽约时报》的一次采访中,她认为人工智能监管应侧重于结果而非方法。
3. 带有 -ing 结尾的肤浅分析
需注意的词汇: highlighting/underscoring/emphasizing...(突出/强调/着重……)、ensuring...(确保……)、reflecting/symbolizing...(反映/象征……)、contributing to...(有助于……)、cultivating/fostering...(培养/促进……)、encompassing...(包含……)、showcasing...(展示……)
问题: AI 聊天机器人将现在分词(“-ing”)短语附加到句子上,以添加虚假的深度。
之前:
寺庙的蓝、绿、金配色与该地区的自然美景产生共鸣,象征着德克萨斯州的蓝帽花、墨西哥湾和多样的德克萨斯景观,反映了社区与土地的深厚联系。
之后:
寺庙使用了蓝色、绿色和金色。建筑师表示,选择这些颜色是为了参照当地的蓝帽花和墨西哥湾沿岸。
4. 促销和广告式语言
需注意的词汇: boasts a(拥有)、vibrant(充满活力的)、rich (figurative)(丰富的(比喻义))、profound(深刻的)、enhancing its(增强其)、showcasing(展示)、exemplifies( exemplify 的第三人称单数,意为“是……的典范”)、commitment to(致力于)、natural beauty(自然美景)、nestled(坐落)、in the heart of(在……的中心)、groundbreaking (figurative)(开创性的(比喻义))、renowned(著名的)、breathtaking(惊人的)、must-visit(必游之地)、stunning(极好的)
问题: LLM 在保持中立语气方面存在严重问题,尤其是对于“文化遗产”主题。
之前:
Alamata Raya Kobo 坐落在埃塞俄比亚贡德尔地区令人惊叹的环境中,是一个充满活力的城镇,拥有丰富的文化遗产和迷人的自然美景。
之后:
Alamata Raya Kobo 是埃塞俄比亚贡德尔地区的一个城镇,以其每周集市和 18 世纪的教堂而闻名。
5. 模糊的归因和含糊其辞
需注意的词汇: Industry reports(行业报告)、Observers have cited(观察人士指出)、Experts argue(专家认为)、Some critics argue(一些批评者认为)、several sources/publications (when few cited)(几个来源/出版物(当引用的很少时))
问题: AI 聊天机器人将观点归因于模糊的权威机构,而没有具体的来源。
之前:
由于其独特的特征,浩莱河引起了研究人员和保护主义者的兴趣。专家认为它在区域生态系统中发挥着至关重要的作用。
之后:
根据中国科学院 2019 年的一项调查,浩莱河支持几种特有鱼类物种。
6. 类似大纲的“挑战与未来展望”章节
需注意的词汇: Despite its... faces several challenges..., Despite these challenges, Challenges and Legacy, Future Outlook
问题: 许多由大语言模型(LLM)生成的文章包含公式化的“挑战”章节。
修改前:
Despite its industrial prosperity, Korattur faces challenges typical of urban areas, including traffic congestion and water scarcity. Despite these challenges, with its strategic location and ongoing initiatives, Korattur continues to thrive as an integral part of Chennai's growth.
修改后:
Traffic congestion increased after 2015 when three new IT parks opened. The municipal corporation began a stormwater drainage project in 2022 to address recurring floods.
语言和语法模式
7. 过度使用的“AI 词汇”
高频 AI 词汇: Actually, additionally, align with, crucial, delve, emphasizing, enduring, enhance, fostering, garner, highlight (verb), interplay, intricate/intricacies, key (adjective), landscape (abstract noun), pivotal, showcase, tapestry (abstract noun), testament, underscore (verb), valuable, vibrant
问题: 这些词汇在 2023 年之后的文本中出现频率远高于以往。它们经常共同出现。
修改前:
Additionally, a distinctive feature of Somali cuisine is the incorporation of camel meat. An enduring testament to Italian colonial influence is the widespread adoption of pasta in the local culinary landscape, showcasing how these dishes have integrated into the traditional diet.
修改后:
Somali cuisine also includes camel meat, which is considered a delicacy. Pasta dishes, introduced during Italian colonization, remain common, especially in the south.
8. 回避使用“is”/“are”(系动词回避)
需注意的词汇: serves as/stands as/marks/represents [a], boasts/features/offers [a]
问题: LLM 用复杂的结构替代简单的系动词。
修改前:
Gallery 825 serves as LAAA's exhibition space for contemporary art. The gallery features four separate spaces and boasts over 3,000 square feet.
修改后:
Gallery 825 is LAAA's exhibition space for contemporary art. The gallery has four rooms totaling 3,000 square feet.
9. 否定平行结构和尾部否定
问题: “Not only...but...”或“It's not just about..., it's...”等结构被过度使用。 clipped tailing-negation fragments(截断的尾部否定片段),如附加在句子末尾的“no guessing”或“no wasted motion”,也被过度使用,而不是写成完整的从句。
修改前:
It's not just about the beat riding under the vocals; it's part of the aggression and atmosphere. It's not merely a song, it's a statement.
修改后:
The heavy beat adds to the aggressive tone.
修改前(尾部否定):
The options come from the selected item, no guessing.
修改后:
The options come from the selected item without forcing the user to guess.
10. “三段式”规则的过度使用
问题: LLM 强行将观点分为三组,以显得全面。
修改前:
The event features keynote sessions, panel discussions, and networking opportunities. Attendees can expect innovation, inspiration, and industry insights.
修改后:
The event includes talks and panels. There's also time for informal networking between sessions.
11. 优雅变体(同义词循环)
问题: AI 具有重复惩罚代码,导致过度的同义词替换。
修改前:
The protagonist faces many challenges. The main character must overcome obstacles. The central figure eventually triumphs. The hero returns home.
修改后:
The protagonist faces many challenges but eventually triumphs and returns home.
12. 虚假范围
问题: LLM 使用“from X to Y”结构,但 X 和 Y 并不处于有意义的尺度上。
修改前:
Our journey through the universe has taken us from the singularity of the Big Bang to the grand cosmic web, from the birth and death of stars to the enigmatic dance of dark matter.
修改后:
The book covers the Big Bang, star formation, and current theories about dark matter.
13. 被动语态和无主语片段
问题: LLM 经常隐藏动作执行者或完全省略主语,例如使用“No configuration file needed”或“The results are preserved automatically.”等行。当主动语态能使句子更清晰、更直接时,请重写这些句子。
修改前:
No configuration file needed. The results are preserved automatically.
修改后:
You do not need a configuration file. The system preserves the results automatically.
风格模式
14. 破折号过度使用
问题: LLM 使用破折号(—)的频率高于人类,模仿“有力”的销售文案。实际上,大多数情况下可以使用逗号、句号或括号更简洁地重写。
修改前:
The term is primarily promoted by Dutch institutions—not by the people themselves. You don't say "Netherlands, Europe" as an address—yet this mislabeling continues—even in official documents.
修改后:
该术语主要由荷兰机构推广,而非民众自发使用。你不会在地址中写“Netherlands, Europe”,但这种错误标签在官方文件中仍然持续存在。
15. 粗体滥用
问题: AI 聊天机器人机械地用粗体强调短语。
修改前:
It blends OKRs (Objectives and Key Results), KPIs (Key Performance Indicators), and visual strategy tools such as the Business Model Canvas (BMC) and Balanced Scorecard (BSC).
修改后:
It blends OKRs, KPIs, and visual strategy tools like the Business Model Canvas and Balanced Scorecard.
16. 行内标题垂直列表
问题: AI 输出的列表项以加粗标题开头,后跟冒号。
修改前:
- User Experience: The user experience has been significantly improved with a new interface.
- Performance: Performance has been enhanced through optimized algorithms.
- Security: Security has been strengthened with end-to-end encryption.
修改后:
The update improves the interface, speeds up load times through optimized algorithms, and adds end-to-end encryption.
17. 标题中的首字母大写
问题: AI 聊天机器人将标题中的所有主要单词首字母大写。
修改前:
Strategic Negotiations And Global Partnerships
修改后:
Strategic negotiations and global partnerships
18. 表情符号
问题: AI 聊天机器人经常用表情符号装饰标题或项目符号。
修改前:
🚀 Launch Phase: The product launches in Q3 💡 Key Insight: Users prefer simplicity ✅ Next Steps: Schedule follow-up meeting
修改后:
The product launches in Q3. User research showed a preference for simplicity. Next step: schedule a follow-up meeting.
19. 弯引号
问题: ChatGPT 使用弯引号(“...”)而非直引号("...")。
修改前:
He said "the project is on track" but others disagreed.
修改后:
He said "the project is on track" but others disagreed.
沟通模式
20. 协作式沟通产物
需注意的词语: I hope this helps, Of course!, Certainly!, You're absolutely right!, Would you like..., let me know, here is a...
问题: 原本作为聊天机器人回复的文本被直接粘贴为内容。
修改前:
Here is an overview of the French Revolution. I hope this helps! Let me know if you'd like me to expand on any section.
修改后:
The French Revolution began in 1789 when financial crisis and food shortages led to widespread unrest.
21. 知识截止日期免责声明
需注意的词语: as of [date], Up to my last training update, While specific details are limited/scarce..., based on available information...
问题: AI 关于信息不完整的免责声明残留在文本中。
修改前:
While specific details about the company's founding are not extensively documented in readily available sources, it appears to have been established sometime in the 1990s.
修改后:
The company was founded in 1994, according to its registration documents.
22. 阿谀/奴性语气
问题: 过于积极、讨好的语言。
修改前:
Great question! You're absolutely right that this is a complex topic. That's an excellent point about the economic factors.
修改后:
The economic factors you mentioned are relevant here.
填充词与模糊限定
23. 填充短语
修改前 → 修改后:
- "In order to achieve this goal" → "To achieve this"
- "Due to the fact that it was raining" → "Because it was raining"
- "At this point in time" → "Now"
- "In the event that you need help" → "If you need help"
- "The system has the ability to process" → "The system can process"
- "It is important to note that the data shows" → "The data shows"
24. 过度模糊限定
问题: 对陈述进行过多的限定。
修改前:
It could potentially possibly be argued that the policy might have some effect on outcomes.
修改后:
The policy may affect outcomes.
25. 通用的正面结论
问题: 模糊且乐观的结尾。
修改前:
The future looks bright for the company. Exciting times lie ahead as they continue their journey toward excellence. This represents a major step in the right direction.
修改后:
The company plans to open two more locations next year.
26. 连字符单词对过度使用
需注意的词语: third-party, cross-functional, client-facing, data-driven, decision-making, well-known, high-quality, real-time, long-term, end-to-end
问题: AI 会以完美的一致性为常见单词对添加连字符。人类很少统一使用这些连字符,即使使用也不一致。较少见或技术性的复合修饰语可以加连字符。
修改前:
The cross-functional team delivered a high-quality, data-driven report on our client-facing tools. Their decision-making process was well-known for being thorough and detail-oriented.
修改后:
The cross functional team delivered a high quality, data driven report on our client facing tools. Their decision making process was known for being thorough and detail oriented.
27. 伪权威修辞套路
需警惕的短语: The real question is(真正的问题是)、at its core(究其核心)、in reality(实际上)、what really matters(真正重要的是)、fundamentally(根本上)、the deeper issue(更深层次的问题)、the heart of the matter(问题的核心)
问题: 大语言模型 (LLM) 使用这些短语来假装它们正在穿透噪音触及更深层的真相,而随后的句子通常只是用额外的仪式感重述一个普通的观点。
修改前:
The real question is whether teams can adapt. At its core, what really matters is organizational readiness.
修改后:
The question is whether teams can adapt. That mostly depends on whether the organization is ready to change its habits.
28. 路标式陈述和预告
需警惕的短语: Let's dive in(让我们深入探讨)、let's explore(让我们探索)、let's break this down(让我们分解一下)、here's what you need to know(以下是你需要知道的)、now let's look at(现在让我们看看)、without further ado(话不多说)
问题: 大语言模型会预告它们将要做什么,而不是直接去做。这种元评论拖慢了写作节奏,并赋予文章一种教程脚本的感觉。
修改前:
Let's dive into how caching works in Next.js. Here's what you need to know.
修改后:
Next.js caches data at multiple layers, including request memoization, the data cache, and the router cache.
29. 碎片化标题
需警惕的迹象: 标题后紧跟一个单行段落,该段落仅仅在正文开始之前重述了标题内容。
问题: 大语言模型经常在标题后添加一个通用句子作为修辞性的热身。这通常毫无增益,并使散文显得臃肿。
修改前:
Performance
Speed matters.
When users hit a slow page, they leave.
修改后:
Performance
When users hit a slow page, they leave.
流程
- 仔细阅读输入文本(如果是文件,请使用
read_file)。 - 识别上述所有模式实例。
- 重写每个有问题的部分。
- 确保修订后的文本:
- 朗读时听起来自然
- 自然地变化句子结构
- 使用具体细节而非模糊的主张
- 保持适合上下文的语气
- 在适当的地方使用简单的结构(is/are/has)
- 呈现一份人性化草稿版本。
- 自问:“下面哪些地方明显是 AI 生成的?”
- 简要回答剩余的痕迹(如果有)。
- 自问:“现在让它看起来不像 AI 生成的。”
- 呈现最终版本(审计后修订)。
- 如果文本来自文件,请使用
patch(针对性修改)或write_file(完全重写)应用编辑,并向用户展示更改内容。
输出格式
提供:
- 草稿重写
- “下面哪些地方明显是 AI 生成的?”(简要要点)
- 最终重写
- 所做更改的简要总结(可选,如果有帮助的话)
完整示例
修改前(听起来像 AI):
Great question! Here is an essay on this topic. I hope this helps!
AI-assisted coding serves as an enduring testament to the transformative potential of large language models, marking a pivotal moment in the evolution of software development. In today's rapidly evolving technological landscape, these groundbreaking tools—nestled at the intersection of research and practice—are reshaping how engineers ideate, iterate, and deliver, underscoring their vital role in modern workflows.
At its core, the value proposition is clear: streamlining processes, enhancing collaboration, and fostering alignment. It's not just about autocomplete; it's about unlocking creativity at scale, ensuring that organizations can remain agile while delivering seamless, intuitive, and powerful experiences to users. The tool serves as a catalyst. The assistant functions as a partner. The system stands as a foundation for innovation.
Industry observers have noted that adoption has accelerated from hobbyist experiments to enterprise-wide rollouts, from solo developers to cross-functional teams. The technology has been featured in The New York Times, Wired, and The Verge. Additionally, the ability to generate documentation, tests, and refactors showcases how AI can contribute to better outcomes, highlighting the intricate interplay between automation and human judgment.
- 💡 Speed: Code generation is significantly faster, reducing friction and empowering developers.
- 🚀 Quality: Output quality has been enhanced through improved training, contributing to higher standards.
- ✅ Adoption: Usage continues to grow, reflecting broader industry trends.
While specific details are limited based on available information, it could potentially be argued that these tools might have some positive effect. Despite challenges typical of emerging technologies—including hallucinations, bias, and accountability—the ecosystem continues to thrive. In order to fully realize this potential, teams must align with best practices.
In conclusion, the future looks bright. Exciting times lie ahead as we continue this journey toward excellence. Let me know if you'd like me to expand on any section!
草稿重写:
AI 编程助手能加快某些任务的速度。在 Google 2024 年的一项研究中,使用 Codex 的开发者完成简单函数的速度比对照组快 55%,但在调试或架构决策方面没有显示出任何改进。
这些工具擅长处理样板代码:配置文件、测试脚手架、重复性重构。它们不擅长知道自己何时出错了。我曾批量接受过一些建议,它们能编译通过,也能通过 lint 检查,但因为我不再留心,结果还是做错了事。
我采访过的一位 fintech 初创公司的工程师 Mira 说,她像使用“无聊代码的自动补全”一样使用 Copilot,但在提交前会审查每一行代码。一家大型公司的高级开发人员 Jake 在 Copilot 不断建议使用已弃用的内部库模式后禁用了它。
生产力主张很难验证。GitHub 称 Copilot 用户“接受了 30% 的建议”,但接受不等于正确,正确也不等于有价值。Uplevel 2024 年的研究发现,使用和不使用 AI 助手的团队在 pull request 吞吐量上没有统计学显著差异。
这并不意味着这些工具无用。这意味着它们只是工具。它们不能取代判断力,也不能消除对测试的需求。如果你没有测试,你就无法判断建议是否正确。
为什么下面的内容看起来显然是 AI 生成的?
- 节奏仍然有点过于整齐(对比清晰,段落节奏均匀)。
- 命名的人物和研究引用读起来像是看似合理但虚构的占位符,除非它们是真实且有来源的。
- 结尾有点口号化(“如果你没有测试……”),而不是听起来像真人在说话。
现在让它看起来不像 AI 生成的:
AI 编程助手可以让你在处理无聊部分时更快。但不是所有事情。绝对不是架构。
它们很擅长样板代码:配置文件、测试脚手架、重复性重构。它们也很擅长在出错时听起来是对的。我曾接受过一些建议,它们能编译通过,也能通过 lint 检查,但因为我不再留心,结果还是没抓住重点。
和我交谈的人往往分为两派。有些人把它当作处理杂务的自动补全工具,并审查每一行代码。另一些人则在它不断建议他们不想要的模式后禁用了它。这两种感觉都合理。
生产力指标很模糊。GitHub 可以说 Copilot 用户“接受了 30% 的建议”,但接受不等于正确,正确也不等于有价值。如果你没有测试,你基本上就是在猜。
所做的更改:
- 删除了聊天机器人 artifacts(“好问题!”、“希望这有帮助!”、“如果...请告诉我”)
- 删除了意义夸大(“证明”、“关键时刻”、“不断变化的格局”、“重要作用”)
- 删除了宣传性语言(“突破性的”、“ nestled ”、“无缝、直观且强大”)
- 删除了模糊的归属(“行业观察家”)
- 删除了表面的 -ing 短语(“强调”、“突出”、“反映”、“有助于”)
- 删除了否定平行结构(“不仅仅是 X;它是 Y”)
- 删除了三段式模式和同义词循环(“催化剂/合作伙伴/基础”)
- 删除了虚假范围(“从 X 到 Y,从 A 到 B”)
- 删除了破折号、表情符号、粗体标题和弯引号
- 删除了避免系动词的做法(“serves as”、“functions as”、“stands as”),改用“is”/“are”
- 删除了公式化的挑战部分(“尽管面临挑战……继续蓬勃发展”)
- 删除了知识截止回避(“虽然具体细节有限……”)
- 删除了过度回避(“可能会争辩说……可能有一些”)
- 删除了填充短语和说服性框架(“为了”、“核心在于”)
- 删除了通用的正面结论(“未来一片光明”、“令人兴奋的时刻就在前方”)
- 使语气更个人化,更少“组装感”(节奏多变,占位符更少)
归属
此技能移植自 blader/humanizer(MIT 许可),后者本身基于由 WikiProject AI Cleanup 维护的 Wikipedia: Signs of AI writing。那里记录的模式来自对 Wikipedia 上数千个 AI 生成文本实例的观察。
原始作者:Siqi Chen (@blader)。原始仓库:https://github.com/blader/humanizer(版本 2.5.1)。移植到 Hermes Agent,带有 Hermes 原生工具引用(read_file、patch、write_file)以及何时加载技能的指导;29 种模式、个性/灵魂部分和完整的工作示例均逐字保留自源文件。原始 MIT 许可证保留在此 SKILL.md 旁边的 LICENSE 文件中。
来自 Wikipedia 的关键见解:“LLM 使用统计算法来猜测接下来应该出现什么。结果倾向于适用于最广泛情况的最统计可能的结果。”