human-writing
Skill124 starsChanged 8 months agoMITView on GitHub
Write content that sounds natural, conversational, and authentically human - avoiding AI-generated patterns, corporate speak, and generic phrasing
- Installs packages
We split this skill into the rules it gives an agent
14 rules. 0 are common, 13 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 24 edited versions we read keep these.
13
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 292
Use subheadings every 2-3 paragraphs
- Distinctive: line 294
Use bullet lists for 3+ related items
- Distinctive: line 295
Add horizontal rules for major section breaks
- Distinctive: line 296
Use blockquotes for important callouts
- Distinctive: line 342
Would a human say this out loud?: If not, rewrite.
- Distinctive: line 343
Is every claim backed by evidence?: If not, add specifics or remove the claim.
- Very specific: line 344
Could this sentence appear in any other company's blog?: If yes, make it specific to PRPM.
- Distinctive: line 345
Does this assume the reader is dumb?: If yes, trust them more.
- Distinctive: line 346
Am I hedging because I'm uncertain?: If yes, verify facts or own the uncertainty.
- Distinctive: line 347
Is this a transition I can delete?: If yes, delete it.
- Distinctive: line 348
Does this open with enthusiasm instead of information?: If yes, lead with the info.
- Distinctive: line 360
Add a specific number
- Distinctive: line 370
Remove exclamation points (except in code comments where appropriate)
- Distinctive: line 379
If you wouldn't say it in a GitHub issue comment, don't put it in a blog post.
“Common” is measured: kept by at least 6 in 10 of the 24 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Human-Style Writing
This skill helps you write content that reads like it was written by a real person with opinions, personality, and specific knowledge—not a corporate blog generator or AI assistant trying to sound helpful.
Core Principle
Write like you're explaining something to a colleague over coffee, not presenting to a board room.
Good writing is specific, opinionated, and conversational. Bad writing is generic, safe, and sounds like every other tech blog.
What Makes Writing Sound AI-Generated
❌ Patterns to Avoid
1. Over-Enthusiastic Openings
❌ "We're thrilled to announce..."
❌ "Today, we're excited to share..."
❌ "I'm delighted to introduce..."
❌ "Join us on this exciting journey..."
✅ "We built X because Y kept breaking."
✅ "Here's what we learned shipping X to production."
✅ "Most migration tools get you 70% of the way. Here's how we get to 95%."2. Vague Claims Without Evidence
❌ "This revolutionary approach transforms how developers work."
❌ "Leveraging cutting-edge AI technology..."
❌ "A game-changing solution for modern development."
❌ "Unlock the full potential of your workflow."
✅ "Nango used this to migrate 47 repos in 3 days."
✅ "We tested this on Next.js App Router migration—reduced manual fixes from 800 to 40."
✅ "Stripe's migration guide is 12,000 words. This gets it down to 200 lines of code."3. Corporate Buzzword Soup
❌ "Empowering developers to leverage synergies..."
❌ "Best-in-class solutions for enterprise-grade..."
❌ "Seamlessly integrate with your existing ecosystem..."
❌ "Drive innovation through collaborative paradigms..."
✅ "Works with whatever you're already using."
✅ "Detects edge cases your regex won't catch."
✅ "One command. No config file. No surprises."4. Unnecessary Hedging
❌ "This might help you potentially improve..."
❌ "You could possibly consider..."
❌ "This may or may not be useful..."
❌ "Some users have reported that..."
✅ "This cuts migration time in half."
✅ "Use this when codemods aren't enough."
✅ "Three users reported this edge case. We fixed it."5. Generic Transitions
❌ "Let's dive deep into..."
❌ "Without further ado..."
❌ "At the end of the day..."
❌ "It goes without saying..."
✅ Just start the next section. You don't need a transition.
✅ Or use a specific connector: "Here's why that matters:"6. Robotic Lists
❌ "Here are 5 key benefits:
1. Enhanced productivity
2. Improved efficiency
3. Better collaboration
4. Increased flexibility
5. Streamlined workflows"
✅ "This saves you time in three ways:
- No more searching docs for edge cases—they're encoded in the package
- AI applies patterns consistently—you don't chase style violations
- Tests are generated, not written—coverage without the grind"What Makes Writing Sound Human
✅ Patterns to Use
1. Specific Details
❌ "Many developers struggle with migrations."
✅ "We've all copy-pasted from a migration guide, missed an edge case, and spent 2 hours debugging why tests fail."
❌ "Performance is significantly improved."
✅ "Query time dropped from 847ms to 12ms after adding the index."
❌ "Works with popular frameworks."
✅ "Tested on Next.js, Remix, SvelteKit, and Astro."2. Direct, Confident Language
❌ "This approach may help you potentially improve your workflow."
✅ "This cuts your migration time in half."
❌ "You might want to consider using this feature."
✅ "Use this feature when you have more than 10 files to update."
❌ "Some users have found this useful."
✅ "Three teams adopted this last week. All three shipped in under 2 days."3. Honest Limitations
❌ "Our comprehensive solution handles all use cases."
✅ "This won't catch dynamic imports or string templates. You'll need to fix those manually."
❌ "Perfectly seamless migration experience."
✅ "Expect about 5% of edge cases to need manual review. That's down from 30%."
❌ "Works with any codebase."
✅ "Works if you're on TypeScript 4.5+. Earlier versions need a polyfill."4. Conversational Asides
✅ "You could write a 400-line script for this. We did. It broke on Unicode."
✅ "Turns out, most projects have 3-5 edge cases that codemods can't handle."
✅ "We tried docs. Developers don't read them. We tried linting. They ignore the warnings."5. Active Voice, Present Tense
❌ "The package can be installed by running the command."
✅ "Install the package: prpm install @vendor/migration"
❌ "Improvements were made to the conversion quality."
✅ "We improved conversion quality from 78% to 94%."
❌ "It has been observed that users prefer..."
✅ "Users prefer X over Y by a 4:1 margin."6. Strong Opening Sentences
❌ "In this post, we will discuss how migrations work."
✅ "Codemods automate 70% of migrations. This gets you to 95%."
❌ "Let me tell you about a new feature we've added."
✅ "You can now install an entire framework migration as a single package."
❌ "Today, I want to talk about our vision for the future."
✅ "Package managers changed how we ship code. We're doing the same for AI instructions."Tone Calibration
Technical Posts
- Voice: Knowledgeable peer, not teacher
- Assumptions: Reader knows basics, wants specifics
- Evidence: Code examples, performance numbers, real packages
- Length: As long as needed to be complete, as short as possible to respect time
Example:
# Converting Copilot Rules to Claude Format
GitHub Copilot uses a single `.github/copilot-instructions.md` file with YAML frontmatter. Claude uses separate skills in `.claude/skills/`.
Here's how we handle the conversion:
1. Parse the YAML frontmatter with js-yaml
2. Extract the `applyTo` glob patterns
3. Convert to Claude's `fileMatch` format
4. Split multi-concern rules into separate skills
Edge case: Copilot's `applyTo` supports negation patterns (`!**/*.test.ts`). Claude doesn't. We preserve these as comments and log a warning.
Conversion quality: 94% (6% requires manual review for negation patterns).Vision Posts
- Voice: Opinionated builder with receipts
- Assumptions: Reader is skeptical, needs convincing
- Evidence: Real-world examples, before/after, objections addressed
- Length: Long enough to make the case, tight enough to stay focused
Example:
# Why Docs Aren't Enough
Stripe's migration guide is 12,000 words. It's comprehensive, well-written, and most developers skim it.
Why? Because reading docs requires:
1. Find the right section (3-5 minutes)
2. Understand the pattern (5-10 minutes)
3. Apply to your specific case (10-30 minutes)
4. Repeat 20-50 times per migration
That's 6-15 hours. And you'll still miss edge cases.
PRPM packages encode those patterns once. AI applies them consistently. Total time: 20 minutes.Tutorial Posts
- Voice: Experienced guide who's made the mistakes
- Assumptions: Reader wants to follow along, copy/paste, learn
- Evidence: Runnable examples, expected output, common pitfalls
- Length: Complete walkthrough with no missing steps
Example:
# Publishing Your First PRPM Package
## What You'll Build
A Cursor rule that enforces "no default exports" across your TypeScript codebase. By the end, you'll have published it to the registry.
## Prerequisites
- Node.js 18+ (check: `node --version`)
- PRPM CLI installed (`npm install -g prpm`)
- GitHub account (for publishing)
## Step 1: Initialize the Package
```bash
$ mkdir no-default-exports
$ cd no-default-exports
$ prpm init
Format: cursor
Subtype: rule
Name: no-default-exports
Description: Enforce named exports in TypeScriptThis creates prpm.json and .cursorrules.
Step 2: Write the Rule
Edit .cursorrules:
[... full example ...]
## Structural Techniques
### 1. Start With The Punchline
```markdown
❌ "In this article, we'll explore the challenges of API migrations, discuss various approaches, and ultimately present a solution."
✅ "API migrations fail because docs explain the 'what' but not the 'why.' Here's how to ship the reasoning with the code."2. Use Subheadings as Scannable Statements
❌ ## Introduction
❌ ## Background
❌ ## Methodology
❌ ## Results
✅ ## The Problem: Docs Go Stale
✅ ## Why Codemods Aren't Enough
✅ ## What PRPM Packages Add
✅ ## Real Example: Next.js App Router3. Show, Don't Just Tell
❌ "The conversion process is simple and efficient."
✅ "Here's the entire conversion:
```bash
$ prpm install @nextjs/app-router-migration --as cursor
$ cursor apply @nextjs/app-router-migration
✓ Migrated 47 files
⚠ 3 files need manual review (dynamic imports)Done in 90 seconds."
### 4. Break Up Walls of Text
- Use subheadings every 2-3 paragraphs
- Insert code blocks to give eyes a break
- Use bullet lists for 3+ related items
- Add horizontal rules for major section breaks
- Use blockquotes for important callouts
### 5. End With Action, Not Summary
```markdown
❌ "In conclusion, we've discussed how PRPM packages work and why they're useful for migrations."
✅ "Try it:
```bash
prpm install @popular/package-nameHave questions? Follow @prpmdev or open an issue."
## Voice Examples from PRPM
### Good (from VISION.md):
> "Codemods automate the first 60–80% of migrations. Docs explain the rest. Developers still wrestle with edge cases, conventions, and tests."
**Why it works:** Specific percentages, clear problem statement, no fluff.
> "You could read 47 migration guides. Or install one package."
**Why it works:** Concrete number, stark contrast, confident.
> "We tried this on Nango's SDK migration. 47 repos, 3 days, zero regressions."
**Why it works:** Real company, real numbers, honest outcome.
### Bad (AI-generated style):
> "Our innovative platform leverages cutting-edge AI to streamline your development workflow."
**Why it fails:** Buzzwords, vague, could describe anything.
> "We're excited to announce a revolutionary new approach to migrations."
**Why it fails:** Over-enthusiastic, no specifics, marketing speak.
> "This powerful solution empowers teams to unlock their full potential."
**Why it fails:** Empty claims, corporate jargon, meaningless.
## Self-Check Questions
Before publishing, ask:
1. **Would a human say this out loud?** If not, rewrite.
2. **Is every claim backed by evidence?** If not, add specifics or remove the claim.
3. **Could this sentence appear in any other company's blog?** If yes, make it specific to PRPM.
4. **Does this assume the reader is dumb?** If yes, trust them more.
5. **Am I hedging because I'm uncertain?** If yes, verify facts or own the uncertainty.
6. **Is this a transition I can delete?** If yes, delete it.
7. **Does this open with enthusiasm instead of information?** If yes, lead with the info.
## Quick Fixes
### If it sounds too formal:
- Replace "utilize" → "use"
- Replace "in order to" → "to"
- Replace "at this point in time" → "now"
- Replace "for the purpose of" → "for" or "to"
- Cut "very," "really," "quite," "actually"
### If it sounds too generic:
- Add a specific number
- Name a real company/project
- Include a code example
- Mention a concrete edge case
- Quote user feedback
### If it sounds too salesy:
- Replace superlatives with comparisons
- Replace "revolutionary" with "different because"
- Replace "amazing" with specific benefits
- Remove exclamation points (except in code comments where appropriate)
- Cut the first paragraph (usually marketing fluff)
## Remember
PRPM users are developers. They have good bullshit detectors. Write like you respect their intelligence and their time.
**Good writing is revision.** First draft: get ideas down. Second draft: cut 30%. Third draft: add specifics. Fourth draft: read it out loud.
If you wouldn't say it in a GitHub issue comment, don't put it in a blog post.Other versions41 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 41
- KKKKhazix/human-writing★ 3.4ksame name
- syahiidkamil/Software-Engineer-AI-Agent-Atlas★ 404same name
- olo-dot-io/Uni-CLI★ 275same name
- Prism-Shadow/deepseek-harness-book★ 124same name
- w93139/human-writing★ 10same name
- yuanhao667/Skills★ 10same name
- jc98629/ai-skills-library★ 9same name
- Moro-w/natural-writing-skill★ 9same name
- codemistake/human-writing★ 4same name
- marcus/marcus-skills★ 4same name
31 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
Used alongsideProjects with it also use leader, storage-analyzer and 3 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
baizhine999Skillleader
AI 扮演你的 Leader,用大厂管理思维审视你的方案、想法和工作,push 你成长。触发词:/leader, leader模式, 大厂模式, 帮我看看这个, 我有个方案, 我最近有点迷茫, 我不知道该怎么做。AI 是施压者,你是被 push 的那个人。
Also in 3 projects that use this
xiaolinVVSkillstorage-analyzer
macOS / Windows 只读存储分析助手(自动识别系统)。扫描整机磁盘占用,找出 占空间大户,把每一项分成 🟢可自动清理 / 🟡需人工判断 / 🔴谨慎清理 三级并给出 可执行处置方案,生成排版精美、可折叠、命令可一键复制的交互式 HTML 报告,并可 起本地服务在网页上一键删除(移废纸篓/直接删)。扫描全程只读。务必在以下场景 使用:用户说"存储分析""磁盘满了""C盘/硬盘满了""空间不够""清理空间" "清理磁盘""占空间""哪些东西占地方""帮我看看存储""看一下电脑存储/空间" "存储空间""电脑空间不够""内存满了/不够/不足""看下内存/存储"(中文口语里 "内存"常指存储空间)"storage analysis""disk cleanup""清缓存""磁盘清理"; 或用户抱怨电脑没空间、想知道什么东西吃硬盘、想要清理建议时。注意:若用户明确 指运行内存/RAM(如"哪个进程吃内存""内存占用高"想看活动监视器),那是 RAM 不是存储,不属于本 skill。
Also in 3 projects that use this
Wyx1105170322Skillneat-freak
End-of-session knowledge cleanup with OCD-level rigor — reconciles project docs (CLAUDE.md, README.md, docs/) and agent memory against the code so nothing rots. 会话结束后对项目文档和记忆进行洁癖级审查与同步。MUST trigger when the user says: "sync up", "tidy up docs", "update memory", "clean up docs", "/sync", "/neat", "同步一下", "整理文档", "整理一下", "更新记忆", "梳理一下", "收尾", "这个阶段做完了", "新人能直接上手", or any phrase suggesting a dev milestone where knowledge needs reconciliation. Also trigger when the user reports stale docs, conflicting memories, or wants a clean handoff to teammates or other agents. Bare "整理" / "tidy" with prior dev context counts — do not under-trigger. Cross-platform: works on Claude Code, OpenAI Codex, OpenCode, and OpenClaw.
Also in 3 projects that use this
VercelSkillfind-skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
Also in 4 projects that use this
SentrySkillskill-creator
Create new agent skills following the Agent Skills specification. Use when asked to "create a skill", "add a new skill", "write a skill", "make a skill", "build a skill", or scaffold a new skill with SKILL.md. Guides through requirements, writing, registration, and verification.
Also in 3 projects that use this
In this repository64 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- beanstalk-deployCursor rule
- .cursor/rules/core-principles.mdcCursor rule
- .cursor/rules/creating-agents-md.mdcCursor rule
- .cursor/rules/creating-claude-agents.mdCursor rule
- .cursor/rules/creating-claude-commands.mdCursor rule
- .cursor/rules/creating-claude-hooks.mdCursor rule
- .cursor/rules/creating-continue-packages.mdCursor rule
- .cursor/rules/creating-copilot-packages.mdCursor rule
- .cursor/rules/creating-cursor-commands.mdCursor rule
- .cursor/rules/creating-cursor-rules.mdcCursor rule
- .cursor/rules/creating-kiro-agents.mdcCursor rule
- .cursor/rules/creating-kiro-packages.mdCursor rule
- .cursor/rules/creating-skills.mdcCursor rule
- .cursor/rules/creating-windsurf-packages.mdCursor rule
- .cursor/rules/elastic-beanstalk-deployment.mdCursor rule
- .cursor/rules/format-conversion.mdcCursor rule
- .cursor/rules/github-actions-testing.mdcCursor rule
- .cursor/rules/karen-repo-reviewer.mdcCursor rule
- .cursor/rules/prpm-json-best-practices.mdcCursor rule
- .cursor/rules/self-improve-cursor.mdcCursor rule
- .cursor/rules/testing-patterns.mdcCursor rule
- .cursor/rules/typescript-type-safety.mdcCursor rule
- .cursor/rules/typescript-type-specialist.mdcCursor rule
- creating-agent-skillsSkill
- adding-new-ai-formatSkill
- agent-builderSkill
- aws-beanstalk-expertSkill
- beanstalk-deploySkill
- claude-hook-writerSkill
- creating-agents-mdSkill
- creating-claude-agentsSkill
- creating-claude-commandsSkill
- creating-claude-hooksSkill
- creating-claude-rulesSkill
- creating-continue-packagesSkill
- creating-copilot-packagesSkill
- creating-cursor-commandsSkill
- creating-cursor-rules-skillSkill
- creating-kiro-agentsSkill
- creating-kiro-packagesSkill
- creating-opencode-agentsSkill
- creating-opencode-pluginsSkill
- creating-skillsSkill
- creating-windsurf-packagesSkill
- creating-zed-extensionsSkill
- documentation-standardsSkill
- elastic-beanstalk-deploymentSkill
- github-actions-testingSkill
- integrating-stripe-webhooksSkill
- karen-repo-reviewerSkill
- npm-trusted-publishingSkill
- osgrepSkill
- postgres-migrationsSkill
- prpm-developmentSkill
- prpm-json-best-practicesSkill
- pulumi-troubleshootingSkill
- self-improvingSkill
- slash-command-builderSkill
- The first 60 shown.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

