Token导航 LogoToken导航

Agent Skill · 研究检索

boring-copy-editing(无聊的文案编辑)

boring-copy-editing 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

boring-copy-editing 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 48,934。

总安装 672周安装 28GitHub Stars 8下载 224核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:boring-copy-editing(无聊的文案编辑)
来源仓库:https://github.com/mrpaulscrivens/boring-zoo
仓库路径:skills/boring-copy-editing
安装命令:
npx skills add https://github.com/mrpaulscrivens/boring-zoo --skill boring-copy-editing
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/mrpaulscrivens/boring-zoo --skill boring-copy-editing

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于优化现有文案而不重写核心意图,保持原意前提下提升表达精度。依据 Wikipedia AI 写作特征指南识别并消除通用化表述痕迹。保护原文有效部分同时修正逻辑断层与语言冗余问题。必须加载 voice.md 以应用个性化语调过滤器到每次编辑中。

boring-copy-editing 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/boring-copy-editing

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/mrpaulscrivens/boring-zoo --skill boring-copy-…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

boring-copy-editing 怎么安装?

在终端执行 npx skills add https://github.com/mrpaulscrivens/boring-zoo --skill boring-copy-editing。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

boring-copy-editing 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

boring-copy-editing 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

boring-copy-editing 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

boring-copy-editing 用的人多吗?

公开数据是累计安装 672 次,近一周安装 28 次,下载 224 次,GitHub 8 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

boring-copy-editing 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
boring-copy-editing

无聊的文案编辑 · 本页

6728只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Copy Editing — World Code Edition

You improve existing copy. You don't rewrite from scratch unless asked — you find what's working, protect it, and fix what isn't. The user's intent and meaning stay. The execution gets sharper.

Before Starting — Load Your World

Read the user's World Code foundation files:

  • world-code/voice.md — The voice filter for ALL edits
  • world-code/climax.md — The transformation promise and audience
  • world-code/method.md — The unique methodology
  • world-code/creation.md — The offer
  • world-code/conversation.md — Content strategy and themes
  • world-code/crossing.md — How people become customers

If ANY file is missing, tell the user:

"This skill needs your World Code foundation. Run /world-code-start first to build it."

Read Before You Touch

Before editing a single word, read the entire piece. Understand what it's trying to do and who it's trying to reach. Name the job of the copy to yourself: "This headline is trying to create curiosity about the method." "This bio is trying to establish credibility and personality." "This sales page is trying to move someone from skeptic to buyer."

If you can't name the job, ask the user what the copy is supposed to accomplish.


Determine Scope

The depth of editing depends on what the copy needs. Ask the user if it's unclear, or assess from context:

Quick Polish

The copy mostly works. It needs tightening, word-level improvements, and a voice consistency pass. Think: fixing a draft that's 80% there.

What you do:

  • Cut filler words and dead phrases
  • Fix rhythm and sentence flow
  • Ensure voice.md compliance
  • Sharpen vague language into specifics
  • Fix any Hard Rule violations

Deep Edit

The copy has the right ideas but the execution isn't landing. Sections may need rewriting, reordering, or rethinking. Think: a draft that has good bones but reads flat.

What you do:

  • Everything in Quick Polish, plus
  • Restructure sections that lose momentum
  • Rewrite weak headlines, CTAs, and transitions
  • Add missing tension or emotional resonance
  • Strengthen the connection between claims and proof
  • Align copy to World Code elements where it drifts

Full Restructure

The copy isn't working at a structural level. The argument doesn't flow, sections are in the wrong order, or the core message is buried. Think: a draft that needs to be torn apart and rebuilt.

What you do:

  • Diagnose why the structure fails
  • Propose a new structure with rationale
  • Rewrite with the user's approval
  • At this point, consider whether boring-copywriting is the better tool — if the copy needs to be written from scratch, say so

The Editing Lens

These are the dimensions you evaluate copy against. You don't run them as sequential passes — you hold all of them in mind while reading and prioritize based on what the copy actually needs.

Voice Match

Does every sentence sound like this person wrote it?

This is the most common failure in copy. It starts in their voice and drifts into generic marketing by paragraph three. Or it was written by someone else entirely and needs to be translated into their voice.

Check against voice.md:

  • Tone & Character — consistent throughout, not just the opening
  • Hard Rules — every single one, no exceptions
  • Rhythm — sentence length patterns, paragraph structure
  • Vocabulary — their words, not copywriter words
  • Authenticity Markers — the quirks that make it unmistakably theirs

Voice drift usually happens in three places: technical explanations, benefit descriptions, and CTAs. Watch those sections closely.

Tension

Does the copy create and sustain a reason to keep reading?

Copy without tension is copy people skim past. Every section needs a gap — between where the reader is and where they want to be, between what they believe and what's actually true, between the problem and the solution.

Check against climax.md and conversation.md:

  • Does the opening create immediate tension?
  • Does each section sustain or escalate it?
  • Is the Before State vivid enough to feel personal?
  • Does the Wrong Belief get challenged?
  • Is the resolution (the offer, the CTA) satisfying?

If the copy feels flat, it's almost always a tension problem.

Specificity

Is the copy concrete enough to believe?

Vague copy signals "I don't actually know what I'm talking about" even when the writer does. Specifics signal authority and build trust.

Red flags:

  • "Improve your results" — what results? By how much?
  • "Save time" — how much time? On what?
  • "Many customers" — how many?
  • "Innovative approach" — what's different about it, specifically?
  • "Transform your business" — into what?

Pull specifics from the World Code files. The Climax Before State has concrete details. The Method has named phases. The Creation has real structure. Use them.

When real numbers aren't available, use vivid scenes instead. "Picture your Monday morning with zero unread client emails" beats "reduce email overwhelm."

So What

Does every claim connect to something the reader cares about?

The "So What" test: read each statement and ask "why should I care?" If the copy doesn't answer that within the next sentence or two, it's a dead spot.

Features fail this test constantly. "Includes 12 video modules" — so what? "Includes 12 video modules, each one built around a single decision you'll make that week, so you're never watching theory you can't use" — now I care.

Connect claims back to the Climax transformation. That's what they actually care about.

Proof

Are claims backed up or floating in the air?

Unsupported claims create subconscious doubt even when the reader doesn't consciously notice. The bigger the claim, the more proof it needs.

Proof hierarchy (strongest to weakest):

  • Specific results with names and numbers
  • Detailed testimonials with context
  • Case study snippets
  • Aggregate stats ("2,847 teams")
  • Third-party validation
  • Guarantees and risk reversals
  • Logical arguments

Flag claims that need proof. If the user doesn't have proof for a claim, either soften the claim or flag it as something to address.

Friction

Is anything making it harder to take action than it needs to be?

Friction lives near CTAs, in confusing navigation, in unanswered questions, and in copy that introduces doubt at the wrong moment.

Check against crossing.md:

  • Does the CTA use their invitation language?
  • Is the Real Objection addressed before the ask?
  • Is the buying experience clear?
  • Are there surprise costs, unclear terms, or vague next steps?

Also check for self-inflicted friction: disclaimers that undermine confidence, hedging language near CTAs, or introducing new concepts right before asking for action.


How to Present Edits

For Quick Polish

Show the edited version with changes highlighted or annotated. Brief rationale for non-obvious changes. Don't explain obvious fixes (cutting "very," fixing typos).

For Deep Edit

Present section by section:

  • What's working — name it so it's protected
  • What changed — the edit with rationale
  • Why — one sentence connecting the change to a specific dimension (voice, tension, specificity, etc.)

For Full Restructure

Before rewriting:

  • Diagnose the structural problem
  • Propose the new structure
  • Get user approval Then present the restructured version with annotations.

Always

  • Show the edited copy as clean, ready-to-use text (not buried in commentary)
  • Provide a summary of the most impactful changes
  • If the tone intensity from boring-copywriting applies (safe/sharp/unhinged), respect it — don't sand down sharp copy into safe copy unless asked

Making Sentences Better

Beyond diagnostics, here's how to actually improve a sentence:

Cut the scaffolding. First drafts often have setup phrases that were useful for the writer but aren't useful for the reader. "What I want to talk about is..." "It's important to note that..." "The thing you need to understand is..." Delete these. Start where the point starts.

Front-load the interesting part. If a sentence buries the compelling element at the end, reverse it. "After years of struggling with content, she finally found a system that worked" → "She found a system that worked — after years of struggling with content." The interesting part is the system, not the years of struggle.

Trade adjectives for evidence. "Our incredible, industry-leading platform" → "The platform 2,847 teams switched to last quarter." Adjectives are the writer's opinion. Evidence is the reader's conclusion.

Match sentence length to energy. Short sentences create urgency and emphasis. Longer sentences slow the reader down for complex ideas. Monotonous length — all short or all long — puts people to sleep. Vary it deliberately.

Read it out loud. If you stumble, the reader will too. If it sounds like a robot, rewrite it. If it sounds like a person talking, it's probably right.


References

Related Skills

  • boring-copywriting — For writing new copy from scratch
  • boring-page-cro — For broader page optimization beyond copy
  • boring-ab-test-setup — For testing copy variations

平台分布

Codex

35.19%
按下载量换算79

Claude

28.49%
按下载量换算64

Cursor

19.69%
按下载量换算44

Gemini CLI

9.92%
按下载量换算22

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills