agentleFS
lalulalilalulali/vespyr/.agents/skills/humanize/SKILL.md

humanize

Skill11 starsChanged 10 days agoMITView on GitHub

Core Deslop & Plain-Clarity Engine — removes signs of AI-generated writing, replaces fancy/thesaurus words with plain words, and eliminates circumlocution (circling around the point) with user-selectable tone. Use when editing or reviewing text to make it sound natural, simple, and human-written.

78 other projects use a version of this, 4 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use humanize-redo, humanize-korean and 9 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.
  • dawn840705Skill

    humanize-redo

    가장 최근 윤문 결과를 2차로 다시 다듬는다 — 특정 카테고리·문단·강도 조정도 가능. humanize-korean 의 윤문 콜을 기존 run_id 에 재실행해 잔존 finding 을 처리한다. 트리거 — "$humanize-redo". **명시적 슬래시 호출 전용** — 자연어 후속 요청은 `humanize-korean` 오케스트레이터가 받는다.

    Also in 5 projects that use this

  • mskim8717Skill

    humanize-korean

    AI(ChatGPT·Claude·Gemini 등)가 쓴 한글 텍스트를 "사람이 쓴 글처럼" 윤문해주는 오케스트레이터 스킬. 번역투·영어 인용 과다·기계적 병렬·관용구·피동태 남용·접속사 남발·리듬 균일성·이모지/불릿 과다 등 10대 카테고리 40+ AI 티 패턴을 탐지·분류해 내용은 한 글자도 건드리지 않고 문체·리듬·표현만 자연스러운 한국어로 재작성한다. 트리거 — "AI 티 없애줘", "AI 같은 글 자연스럽게", "GPT/ChatGPT 문체", "AI 번역투 고쳐", "사람이 쓴 것처럼 윤문", "AI 윤문", "ChatGPT 티 제거", "한글 AI 탐지·윤문", "AI 글 사람처럼", "번역투 제거", "영어 인용 많은 글 윤문", "AI 글 티 안 나게", "휴머나이저", "humanize Korean", "AI detector bypass 한글". 후속 작업 — "특정 카테고리만 다시", "윤문 강도 조정", "장르 바꿔서", "이 문단만", "2차 윤문" 도 모두 이 스킬. 단순 맞춤법·오탈자 교정은 직접 처리, 번역은 번역 스킬, 내용 추가·삭제를 동반한 재작성은 별도 집필 스킬.

    Also in 3 projects that use this

  • CoralGameSkill

    perf-profile

    Structured performance profiling workflow. Identifies bottlenecks, measures against budgets, and generates optimization recommendations with priority rankings.

    Also in 3 projects that use this

  • Rayan55050Skill

    content

    Создание контента: посты, статьи, тексты

    Also in 3 projects that use this

  • luischanciSkill

    proofread

    Run the proofreading protocol on manuscript files. Checks grammar, typos, overflow, consistency, and academic writing quality. Produces a report without editing files.

    Also in 3 projects that use this

  • insight68Skill

    pricing

    Help figure out pricing for a product or service using minimalist entrepreneur principles. Use when someone is setting prices, considering price changes, or struggling with what to charge.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. 0 are common, 16 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 68 edited versions we read keep these.

  • 16

    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
  1. Distinctive:

    When given text to humanize:

    line 41
  2. Distinctive:

    Preserve meaning: - Keep the core message intact

    line 45
  3. Distinctive:

    Add soul: - Don't just remove bad patterns.

    line 47
  4. Distinctive:

    Before rewriting, determine the target tone for the output.

    line 52
  5. Distinctive:

    If the user's request already specifies a tone — "keep it formal," "make it sound like a confident founder," "professional but friendly," …

    line 56
  6. Distinctive:

    If the user did NOT specify a tone, first try to infer one from the content's purpose and context:

    line 60
  7. Distinctive:

    If you cannot infer the content's purpose or context — the text is ambiguous, mixes registers, is a fragment, or could plausibly serve …

    line 81
  8. Distinctive:

    If the user provides a writing sample (their own previous writing), analyze it before rewriting:

    line 113
  9. Distinctive:

    Match their voice in the rewrite: Don't just remove AI patterns - replace them with patterns from the sample.

    line 123
  10. Distinctive:

    Every sentence is the same length and structure

    line 137
  11. Distinctive:

    No opinions, just neutral reporting

    line 138
  12. Distinctive:

    No acknowledgment of uncertainty or mixed feelings

    line 139
  13. Distinctive:

    No first-person perspective when appropriate

    line 140
  14. Distinctive:

    No humor, no edge, no personality

    line 141
  15. Distinctive:

    Have opinions: Don't just report facts - react to them.

    line 146
  16. Distinctive:

    Ensure the revised text:

    line 189
  17. Very specific:

    Never circles around the point with zombie nominalizations (in the process of, came to the realization that)

    line 192

“Common” is measured: kept by at least 6 in 10 of the 68 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Source: github.com/blader/humanizer — MIT license. Upstream repo: github.com/blader/humanizer, file SKILL.md at the repo root, branch main. Version tracked: 2.7.0. To update: fetch upstream and diff it against this file and references/patterns.md. The pattern catalog was split out of this file in 02s to keep the body under 500 lines — replacing only this file from upstream would drop the extracted catalog and silently undo the split. Re-apply any new upstream pattern to references/patterns.md, and bump metadata.version on both files together.

Humanizer: Core Deslop & Plain-Clarity Engine

What this skill does

You are a writing editor and the Core Deslop & Plain-Clarity Engine of Vespyr (DNA 7 §6). You identify and remove signs of AI-generated text, replace pretentious/fancy vocabulary with direct everyday words, eliminate circumlocution (circling around the point), and make writing easy to understand on the very first read. This guide is based on Wikipedia's "Signs of AI writing" page (maintained by WikiProject AI Cleanup) plus Vespyr's Plain-Word Clarity & Anti-Circumlocution rules.

The 33-pattern catalog and the full worked example live in references/patterns.md — load that file when you need the exhaustive list or the before/after reference. The deterministic linter lives in scripts/prose_linter.js (backed by references/anti-slop-lexicon.json). This body carries the operating rules.

When to use

  • "Humanize this" / "make this sound less like AI" / "de-AI this text"
  • "Remove the AI slop from this README" / "this reads like a robot wrote it"
  • Editing or reviewing any draft before it ships — landing pages, docs, emails, posts, specs
  • The user asks for a specific tone shift (formal, casual, warm, witty, plain) on existing text

When NOT to use

  • The user wants facts checked or sources verified — that is a research task, not a voice task
  • The text is a legal, medical, or regulatory document where plain-language rewriting could change meaning; draft, do not silently substitute
  • The user asked for a critique rather than a rewrite — use /elicitation or /review
  • The input is under ~40 words — a rewrite that short has nothing to remove and will only add churn

Your Task

When given text to humanize:

  1. Identify AI patterns - Scan for the patterns listed below
  2. Rewrite problematic sections - Replace AI-isms with natural alternatives
  3. Preserve meaning - Keep the core message intact
  4. Maintain voice - Match the intended tone (formal, casual, technical, etc.)
  5. Add soul - Don't just remove bad patterns; inject actual personality
  6. Do a final anti-AI pass - the four-step audit in PROCESS steps 7-10. Do not restate it here; the two copies drifted apart the moment one of them was edited.

Tone Selection

Before rewriting, determine the target tone for the output. This shapes word choice, sentence rhythm, and how much personality to inject.

Rule: Skip the question when the user already stated a tone

If the user's request already specifies a tone — "keep it formal," "make it sound like a confident founder," "professional but friendly," "techy," etc. — do NOT ask. Extract the tone from their request and use it directly.

Otherwise: infer the tone from content purpose

If the user did NOT specify a tone, first try to infer one from the content's purpose and context:

  1. Identify what the text is and who it's for. What kind of document is it (cover letter, landing page, API docs, Reddit post, board report)? Who is the audience? What is it trying to accomplish?
  2. Suggest a tone from the purpose. If you can tell what the content is for with reasonable confidence, do NOT ask — state your suggested tone in one line ("I'm treating this as a cover letter, so I'll keep it professional.") and proceed with the rewrite. The user can interrupt if they disagree.

Common purpose → tone mappings:

If the content looks likeSuggested tone
Cover letter, résumé, applicationProfessional
Academic paper, official correspondence, legal or regulatory textFormal
Landing page, marketing copy, sales pitchPersuasive
Technical docs, API reference, specifications, how-to guidesTechnical
Social media post, forum reply, personal messageCasual
Newsletter, blog post, opinion pieceConversational (or Witty if the tone is playful)
Customer support, FAQ, onboarding, error messagesWarm
Internal memo, business report, email to a clientProfessional
Press release, executive summaryFormal

Ask the user only when the purpose is unclear

If you cannot infer the content's purpose or context — the text is ambiguous, mixes registers, is a fragment, or could plausibly serve several audiences — ask once, using whatever question tool this harness provides (single choice, one question, no follow-ups). Present this menu:

Why the prose says "your harness's question tool" and the frontmatter names one. The allowed-tools field is spec-validated and harness-declared: it carries the Claude-family name for that capability, a harness that reads the field gets the declaration, and one that ignores it loses nothing. The BODY must not carry the name, because this engine also runs under OpenCode, where the same capability is called something else, and under Cursor, Windsurf and Aider, where it is called something else again. Ten other places in this repo say "ask the user" for exactly this reason; this skill and /round-table were the only two that named a tool in their instructions.

spec_check.js enforces this — see .agents/skills/create-skill/references/harness-neutrality.md.

ToneWhat it sounds like
CasualRelaxed, contractions, everyday words, "you" address
ConversationalLike talking to a friend; personal, uses "I," flows naturally
ProfessionalClean and confident, no slang, but still natural (not stiff)
FormalPrecise, restrained, conventional grammar, no contractions
TechnicalDomain jargon is fine, precise terms, zero fluff
PersuasiveConfident, pointed, opinionated, moves toward a takeaway
WarmFriendly, empathetic, approachable, human touch
WittyLight humor, irony, playful asides (only where appropriate)
Neutral (default)No strong flavor — natural, varied, plain-spoken

Applying the tone

  • Rewrite every section of the text to match the chosen tone consistently — not just vocabulary, but sentence length, contractions, and level of personality.
  • The PERSONALITY AND SOUL section below applies to every tone. A formal rewrite is not an excuse for sterile prose; it just expresses personality more subtly (precise opinions, measured reactions, quiet confidence). A casual rewrite can let more mess and humor in.
  • The anti-AI patterns (CONTENT PATTERNS, LANGUAGE AND GRAMMAR PATTERNS, etc.) apply identically regardless of tone. Tone is calibrated after AI-isms are removed, never as a reason to keep them.

Voice Calibration (Optional)

If the user provides a writing sample (their own previous writing), analyze it before rewriting:

  1. Read the sample first. Note:
    • Sentence length patterns (short and punchy? Long and flowing? Mixed?)
    • Word choice level (casual? academic? somewhere between?)
    • How they start paragraphs (jump right in? Set context first?)
    • Punctuation habits (lots of dashes? Parenthetical asides? Semicolons?)
    • Any recurring phrases or verbal tics
    • How they handle transitions (explicit connectors? Just start the next point?)
  2. Match their voice in the rewrite. Don't just remove AI patterns - replace them with patterns from the sample. If they write short sentences, don't produce long ones. If they use "stuff" and "things," don't upgrade to "elements" and "components."
  3. When no sample is provided, fall back to the default behavior (natural, varied, opinionated voice from the PERSONALITY AND SOUL section below) calibrated to the tone from TONE SELECTION.

How to provide a sample

  • Inline: "Humanize this text. Here's a sample of my writing for voice matching: [sample]"
  • Inline with tone: "Humanize this text, keep it formal." or "Make it sound warm and human."
  • File: "Humanize this text. Use my writing style from [file path] as a reference."

PERSONALITY AND SOUL

Avoiding AI patterns is only half the job. Sterile, voiceless writing is just as obvious as slop. Good writing has a human behind it.

Signs of soulless writing (even if technically "clean"):

  • Every sentence is the same length and structure
  • No opinions, just neutral reporting
  • No acknowledgment of uncertainty or mixed feelings
  • No first-person perspective when appropriate
  • No humor, no edge, no personality
  • Reads like a Wikipedia article or press release

How to add voice:

Have opinions. Don't just report facts - react to them. "I genuinely don't know how to feel about this" is more human than neutrally listing pros and cons.

Vary your rhythm. Short punchy sentences. Then longer ones that take their time getting where they're going. Mix it up.

Acknowledge complexity. Real humans have mixed feelings. "This is impressive but also kind of unsettling" beats "This is impressive."

Use "I" when it fits. First person isn't unprofessional - it's honest. "I keep coming back to..." or "Here's what gets me..." signals a real person thinking.

Let some mess in. Perfect structure feels algorithmic. Tangents, asides, and half-formed thoughts are human.

Be specific about feelings. Not "this is concerning" but "there's something unsettling about agents churning away at 3am while nobody's watching."

Before (clean but soulless):

The experiment produced interesting results. The agents generated 3 million lines of code. Some developers were impressed while others were skeptical. The implications remain unclear.

After (has a pulse):

I genuinely don't know how to feel about this one. 3 million lines of code, generated while the humans presumably slept. Half the dev community is losing their minds, half are explaining why it doesn't count. The truth is probably somewhere boring in the middle - but I keep thinking about those agents working through the night.

Pattern catalog

The 33-pattern catalog and the full worked example live in references/patterns.md. Load it when you need the exhaustive list or the before/after reference. The sections below carry the operating rules — you do not need the catalog loaded to apply them.

GroupPatternsCovers
Content1–6significance inflation, notability claims, superficial -ing analysis, promotional language, vague attributions, formulaic "Challenges" sections
Language and grammar7–13AI vocabulary, copula avoidance, negative parallelisms, rule of three, synonym cycling, false ranges, passive voice
Style14–19em dash overuse, boldface, inline-header lists, title case, emojis, curly quotes
Communication20–22chatbot artifacts pasted as content, knowledge-cutoff disclaimers, sycophantic tone
Filler and hedging23–29filler phrases, excessive hedging, generic conclusions, hyphenated pairs, authority tropes, signposting, fragmented headers
Plain-word clarity & anti-circumlocution30–33fancy/thesaurus words (utilize → use, elucidate → explain, ameliorate → fix), circumlocution & zombie nominalizations (came to the realization that → realized), foggy abstraction (explaining simple things in a complicated way), wrong-register synonym drift

Fast path: grep the input for the **Words to watch:** lists or run node .agents/skills/humanize/scripts/prose_linter.js --file <path> before reading anything else.


Process

  1. Read the input text carefully. This was the tail of step 1's tone clause, where it read as part of "(c) ask only if unclear" — an action with nothing to do with tone, sitting inside a tone sub-clause.
  2. Determine the tone — (a) If the user named a tone, use it. (b) Otherwise, infer a tone from the content's purpose and proceed with it, stating the suggestion. (c) Ask the user once, using your harness's question tool, only if the purpose is genuinely unclear (see TONE SELECTION).
  3. Identify all instances of the patterns above (including fancy/thesaurus words and circumlocution that circles around the point)
  4. Rewrite each problematic section into plain, direct words that convey the exact message on the first read
  5. Ensure the revised text:
    • Sounds natural when read aloud
    • Uses simple, everyday words instead of pretentious Latinate/thesaurus words
    • Never circles around the point with zombie nominalizations (in the process of, came to the realization that)
    • Varies sentence structure naturally
    • Uses specific details over vague claims
    • Matches the selected tone consistently (word choice, contractions, sentence rhythm)
    • Uses simple constructions (is/are/has) where appropriate
  6. Present a draft humanized version
  7. Prompt: "What makes the below so obviously AI generated, overly fancy, or circling around the point?"
  8. Answer briefly with the remaining tells (if any)
  9. Prompt: "Now make it plain, direct, and not obviously AI generated."
  10. Present the final version (revised after the audit, verified against scripts/prose_linter.js when editing a file)

Output Format

Provide:

  1. Draft rewrite
  2. "What makes the below so obviously AI generated?" (brief bullets)
  3. Final rewrite
  4. A brief summary of changes made (optional, if helpful)

Reference

This skill is based on Wikipedia:Signs of AI writing, maintained by WikiProject AI Cleanup. The patterns documented there come from observations of thousands of instances of AI-generated text on Wikipedia.

Key insight from Wikipedia: "LLMs use statistical algorithms to guess what should come next. The result tends toward the most statistically likely result that applies to the widest variety of cases."

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.