Open-source humanizer and anti-slop skills, security-graded: make AI-written text, code and UI read like a person made it (blader/humanizer, Humanizer-zh, taste-skill), and detect AI tells. Vet before you install.
Of the 229 Humanizer Skills & AI Slop Detectors here, the best one to start with is taste-skill (rated SAFE by our security scan, ★ 93.1k). Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop
We ran 26 of these skills on 2026-10-07 and 23 ran: each rewrote the same three texts (an English post, a Chinese post, a short story) in a throwaway sandbox, driven by Claude Code (Claude Opus 5.5), judged by gpt-6-astra.
| Skill | Reads human | Layer reached | AI structure removed | Facts | Detector flags | Evidence |
|---|---|---|---|---|---|---|
| sloptrim ★ 212 | 4.5/5 | Changed wording, not structure | 0 | All kept | 6 (0 structure) | Before / after |
| slop-guard ★ 163 | 4/5 | Changed wording, not structure | 0 | All kept | 5 (2 structure) | Before / after |
| anti-slop ★ 4,501 | 4/5 | Changed wording, not structure | 1 | All kept | - | Before / after |
| humanizer-stack ★ 318 | 3.5/5 | Changed structure | 1 | All kept | - | Before / after |
| academic-humanizer ★ 1,781 | 3.5/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| no_ai_slop_writing_rules ★ 694 | 3.5/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| humanizer-skill ★ 264 | 3.3/5 | Changed wording, not structure | 1 | All kept | 38 (8 structure) | Before / after |
| humanizer ★ 54,072 | 3/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| sepia ★ 2,970 Restructured the Chinese post into a numbered list with a firmer summary: more AI-like structure, not less |
3/5 | Changed structure | 0 | All kept | - | Before / after |
| humanize ★ 515 | 3/5 | Changed wording, not structure | 1 | All kept | - | Before / after |
| unslop ★ 153 | 3/5 | Changed wording, not structure | 1 | All kept | - | Before / after |
| Humanizer-zh ★ 18,940 | 3/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| speak-human-tw ★ 1,027 Rewrites into Traditional Chinese (Taiwan usage) |
3/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| De-AI-Prompt-Enhancer-Writer-Booster-SKILL ★ 848 | 3/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| qu-ai-wei ★ 622 | 3/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| stop-slop ★ 17,755 | 3/5 | Changed wording, not structure | 1 | All kept | - | Before / after |
| StealthHumanizer ★ 156 | 2.3/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| patina ★ 363 | 2/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| lieflat-less-ai-tone ★ 2,303 | 2/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| shuorenhua ★ 1,980 | 2/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| AI-Text-Humanizer-App ★ 435 Prepends stock transitions ("Nonetheless, Every morning…"): reads more like AI, not less |
1/5 | Changed wording, not structure | 0 | All kept | - | Before / after |
| vale-ai-tells ★ 115 | - | Detector only | - | - | 22 (2 structure) | Before / after |
| avoid-ai-writing ★ 4,864 | - | Detector only | - | - | 7 (0 structure) | Before / after |
AI-written text gives itself away on two layers. The surface is word choice and punctuation: "delve", "tapestry", em-dashes, stock phrases. Underneath are the choices: spelling out the lesson, tying every thread up neatly, gesturing at "great works" instead of naming one. Most humanizer skills scrub the surface. A 2026 study of 61,608 stories (StoryScope, University of Maryland and Google DeepMind, COLM 2026) found that this barely matters: after a rewriting tool removed surface artifacts from AI stories, a classifier that looks only at narrative structure still caught 93.9% of them, against 95.5% before. So we judge a humanizer by which layer it reaches.
| Layer | Typical tells | What the evidence says | What scrubbing it leaves |
|---|---|---|---|
| Words and punctuation | "delve", "tapestry", em-dashes, stock openers and closers. | Fading on its own: newer models use far fewer em-dashes, and fine-tuning on human style cut AI detection on creative writing from 97% to 3%. | Easy to fix and easy to fake; fixing it changes nothing else. |
| Phrasing | Clichés, purple prose, redundant exposition. | In the study, a tool rewrote seven kinds of such artifacts; structure-based detection fell only from 95.5% to 93.9%. | The text reads smoother but makes the same choices. |
| Structure | States its own theme (77% of AI stories vs 52% of human ones), resolves into acceptance (47% vs 27%), no subplots (79% vs 57%), emotion through bodily sensation (81% vs 38%), vague allusions instead of named works (72% vs 50%), rarely addresses the reader (7% vs 28%). | Narrative choices alone separate human from AI fiction at 93.2% macro-F1, with every style signal removed. | This is the layer that persists. Changing it takes a real rewrite: drop the stated moral, leave an ending open, name the specific thing. |
The study measured fiction. For posts, emails, reports and docs we apply the same split as a working rule: we have not measured it on non-fiction. The page also lists code and UI tools; there the surface is comment bloat or a default gradient and the structure is what the code or the screen actually does, which is our analogy, not the study's.
| Your own draft, so it stops sounding generic | A phrasing-level humanizer plus a style-rule file you keep |
| Long-form writing you will publish | A tool that reaches structure (drops stated morals, keeps endings open, asks for specifics), then a human edit |
| Chinese text | A Chinese-first tool; English word lists do not carry over |
| Checking text, not rewriting it | A slop detector, read as a list of things to look at, not a verdict |
The floor is install safety. 1 entry on this page is flagged CAUTION or UNSAFE. Check before installing: anti-slop-python (eval_usage).
Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop
Agent skill that removes signs of AI-generated writing from text
```bash
mkdir -p ~/.claude/skills
git clone https://github.com/blader/humanizer.git ~/.claude/skills/humanizer
```Chinese version of Humanizer, a Claude Code skill intended to remove signs that text was generated by AI.
AI writing skill that eliminates detectable AI patterns. Works with Claude Code, Codex, Cursor, Gemini CLI, and 8+ other agents.
A Claude Code skill that removes signs of AI-generated writing from academic medical papers, making them sound more natural and professionally written.
```bash
mkdir -p ~/.claude/skills
git clone https://github.com/matsuikentaro1/humanizer_academic.git ~/.claude/skills/humanizer_academic
```A skill file for removing AI tells from prose
AI가 쓴 한글을 사람 글처럼 윤문하는 Claude 스킬 — Korean AI-text humanizer: detects and rewrites translationese, mechanical parallelism, and 71 other AI tells
Skill that audits and rewrites content to remove AI writing patterns. Use it with your favorite agents including Claude Code, OpenClaw, Codex, and Hermes.
Rules for an AI coding agent to filter out generic AI-generated UI designs, text, and code.
Opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns
Open-source text humanization pipeline with every intermediate step published. Two LLM rewrites at temp 1.3, then two hops across different NMT engines. Four documented methodologies you can read, modify, and run locally.
De-AI writing skill for any Agent Skills-compatible agent (77+ via the Skills CLI), with native plugins for Claude Code, Codex, Grok Build, and Antigravity. Narrative-architecture repair for fiction, venue-matched rules for professional prose. Based on StoryScope (arXiv:2604.03136).
一个基于 283 万字语料统计的去 AI 味 skill · An AI-tone removal skill grounded in a 2.83-million-character corpus study
AIWriteX: WeChat tool for trend analysis, topic ideas, article collection, formatted publishing, AI images, rewriting, multi-platform posts, and video copy.
说人话｜中文优先的去 AI 味改写 skill：保事实、分场景、改完可直接发。Chinese-first rewrite skill for Codex / Claude Code / Cursor / ChatGPT — removes AI tone, preserves facts.
A field guide to the visual & copy tics of AI-generated products — and an Agent Skill that scans your project and strips them out. https://killaislop.com
“Speak Human”: A Traditional Chinese rewriting skill that detects 38 AI writing patterns and fixes Mainland Chinese terms and half-width punctuation for Claude…
The Claude Code skill that directs a website like a film. Commit-sheet, generated assets, build, and an executable anti-slop linter that gates every ship.
VibeCurb - Injects taste into AI workflows to stop agents from generating boring, AI slop.
Make your agents talk human. ADHD-friendly output styles for Claude Code, Codex, and others. So you can pay attention, not tokens.
Strip AI-writing tells from papers and grant proposals (NSF/NIH), while keeping scholarly voice and tying claims to evidence. A skill for Claude Code, Codex, and MorphMind.
Prompts to remove AI-like wording—writer enhancement skill
A GitHub action that detects and automatically closes low-quality and AI slop PRs.
Claude Code reference: write in Louis Rossmann's voice, never like AI slop. Portable CLAUDE.md plus skills.
偏执型设计顾问：反 AI 味设计 + 风格试衣间 + 58 站设计系统库的 Claude Code Skill | Opinionated design advisor for Claude Code: anti-generic UI, style fitting room, 58 real-site design systems.
Remove AI writing traces from Simplified Chinese / Chinese humanizer skill
Best static AI text humanizer. Two research-grounded LLM-agnostic skills that make AI writing sound human and relatable. Nine levers, 50+ peer-reviewed sources, 2024-2026 detection literature.
Transform AI-generated text into formal, human-like, and academic writing with ease, avoids AI detector!
A Claude Code skill that grounds AI-generated UI in real reference images and a persistent per-developer taste profile, instead of generic AI-slop defaults.
humanizer-ru: редактура русского текста для AI-агентов. Убирает канцелярит и шаблонные обороты, сверяет смысл и факты. 67 признаков, 21 жёсткий запрет, сканер, CLI, MCP и расширение Chrome. | Russian text editing for AI agents: removes bureaucratese and template phrasing, checks facts.
De-AI-ify scientific writing
```sh
git clone https://github.com/stephenturner/skill-deslop.git
cd skill-deslop
zip -r deslop.skill SKILL.md references/
```A skill for official writing, based on hundreds of newspaper articles, generating structured, specific, natural documents from type, context, and materials.
AI-writing humanizer for KO/EN/ZH/JA
Design engineering system for AI coding agents — ship UI with craft-level quality. Install as an agent skill.
🎓 降低中文学术写作 AIGC 检测率的 Claude Code Skill | Reduce AIGC detection rate for Chinese academic writing
Eleven free Claude skills that run a LinkedIn account: posts off 21 hook formulas, comments, replies, profile score, weekly plan, and a humanizer that strips the AI fingerprint and scores the draft before it goes out.
Russian text quality for AI agents — neuroslop cleanup, typography, information style, editorial standards, UX writing, business correspondence. Scores text 0–10. 2,000+ linguistic atoms.
Two-pass pipeline for removing AI writing tells from outward-facing text: a surface pass plus a structural pass grounded in the StoryScope study. Packaged as Claude Code Skills. Free community: skool.com/jens-ai-community-1306
Free, open-source AI writing humanizer and detector. 55 patterns, 5 voices, a 0-100 AI-tell score, and nothing leaves your machine.
A local detector for AI-writing patterns. Scores every prose file your agent saves. Python standard library only, no network, no model.
humanizer-zh is a Chinese skill for Codex, Claude Code, and OpenClaw. It edits or reviews long texts to sound native, not translated or AI-made.
Skill для AI-агентов (Claude Code, Codex, OpenClaw, Hermes): убирает 37 признаков AI-генерации из русского текста и проверяет, писала ли его нейросеть - канцелярит, штампы ChatGPT, артефакты копипаста | Russian AI-text humanizer & detector skill for coding agents
Deutscher Text-Humanizer für Claude Code & Codex. Redigiert KI-Floskeln, schützt deine Schreibstimme und prüft 72 deutsche Muster quellentreu.
去除中文文本中 AI 写作痕迹的 Agent Skill（基于 blader/humanizer 与 op7418/humanizer-zh）
Slop Scoring to Stop Slop
🔓 Free open-source AI text humanizer — bypass GPTZero, Turnitin & AI detectors with 16+ Languages support. 35 providers, 4 rewrite levels, 6 Writing Styles, 9 Text Purposes, 13 Tone Presets, multi-pass ninja mode. No login. Built for students & writers.
Make AI output sound human. Strips AI-isms (sycophancy, stock vocab, hedging stacks, em-dash pileups), preserves code/URLs/headings. Plugin for Claude Code, Cursor, Windsurf, Codex, Cline, Copilot, Gemini.
A Claude Code skill that turns AI-sounding Japanese into human-written text. Includes a 20-pattern checklist and rewriting guide.
Static analysis for the debt AI agents leave in PHP: 26 rules, Claude Code hooks, git-diff review, a Rector and Pint fix pass, Pest expectations, CI annotations and an MCP server. Deterministic, local, no LLM.
Stop your AI from writing like AI. A universal system prompt eliminating every known LLM style tell — works with Claude Code, Gemini CLI, Codex CLI, Copilot, Cursor, and any web AI.
Deletion-first Agent Skill for removing test bloat, verification theater, and speculative fallbacks while preserving behavior.
OpenClaw skill that detects and removes signs of AI-generated writing, making text sound natural and human. Based on Wikipedia's Signs of AI Writing.
Chinese-English academic text editing and humanization skill that removes signs of AI generation while strictly preserving factual accuracy.
X (Twitter) marketing skills for Claude Code and Codex: write tweets, threads, and replies in your voice, strip AI tells, and publish via Publora. Open source, MIT. Content engineering by Creative Content Crafts.
In today's rapidly evolving landscape, vale-ai-tells is a comprehensive, cutting-edge Vale style package that empowers writers to seamlessly delve into the rich tapestry of AI tells. It's not just a rule set; it's a game-changer that supercharges your prose, unlocks new possibilities, and really lands. Ship cleaner prose. Full stop. 🚀
```ini
StylesPath = styles
MinAlertLevel = suggestion
Packages = https://github.com/tbhb/vale-ai-tells/releases/download/v1.19.0/ai-tells.zip, \
https://github.com/tbhb/vale-ai-tells/releases/download/v1.19.0/ai-tells-commits.zip
[*.md]
BasedOnStyles = ai-tells
```Quality gate for AI/Codex-generated pull requests: blocks TODO leftovers, leaked secrets, sloppy commits and red CI before they reach main.
Anti-slop skills for AI agents: slopbeth cleans the shipped writing, slopgent cleans the conversation.
Web novel writing skill package covering ranking scans, text analysis, writing, removing AI-like language, and cover image creation for long and short novels.
Codex skill for Huawei Cup papers: detects AI traces, diagnoses sentences, rewrites templates, improves charts, checks rules, and audits submissions.
Stop shipping AI slop. Detects empty functions, fake documentation, and inflated comments in AI-generated code. Production-ready.
An agent skill with multiple gates to check and to remove any AI slop. Use it with codex/claude code or any of your favorite AI tools.
Skill para Claude Code que remove marcas típicas de textos gerados por IA, deixando a escrita mais natural, clara e próxima da linguagem humana. A skill analisa padrões comuns de escrita de modelos de linguagem e reescreve o texto para eliminar sinais artificiais, melhorar fluidez e aumentar naturalidade.
A comprehensive web application that detects AI-generated content in PDF documents and transforms AI text into natural human-like writing. Built with Streamlit, spaCy, and Hugging Face transformers.
Deterministic design critique for HTML/CSS: 73 slop guards from the rules professional designers use in production-grade designs. Exact detectors, idempotent auto-fixes, the anti-slop skill, and the /gesso-critique command. By Gesso, the AI creative team for builders.
Python AI Slop Detector - Find over-engineering, hallucinations, and dead code in Python codebases
Claude Skill removes Chinese AI prose: parallel triplets, nominalizations, abstract subjects; uses concrete details. From hardikpandya/stop-slop.
All slop must die. Airlock is where every git push turns into a slop-free PR.
Transform AI-generated text into natural human-like content. 100% offline · 25 languages · Zero dependencies · PHANTOM™ · ASH™ · Python/PHP/TypeScript
A Codex Skill that removes AI-like phrasing from Chinese and improves style, drawing on Mengke / 好事 to make Chinese drafts sound more human.
Audits AI-generated frontend and rewrites it so it stops looking AI-made: purple-gradient slop and the "tasteful" defaults (cream + terracotta, mono chrome). Zero-dependency scanner included. The design counterpart to avoid-ai-writing.
Design intelligence engine for AI coding tools (Claude Code, Cursor, Windsurf). Deterministic anti-AI-slop linter with 152 rules, 160 brand specs, a 7-axis synthesizer, an MCP server (18 tools), 25 commands, 17 IDEs. Offline, never calls an LLM. MIT.
English text humanizer for Claude: typography, vocabulary, structure. Calibrates to your voice. Built on the UMD / Google DeepMind study and Wikipedia's Signs of AI writing
Coding agent behavior plugin repository: humanize adds “write like a human” rules; skill-stats tracks skill triggers and sources.
humanizer, but for code — an agent skill that removes AI-generated code slop: duplicated helpers, try-import fallbacks, broad excepts, speculative abstractions. Test-gated, behavior-preserving.
```bash
npx skills add LeonardNJU/code-humanizer
```A Claude skill for removing AI traces from Chinese writing. Combines three rule sets, tested on corpora, and resolves seven conflicts between them.
Türkçe metinlerden yapay zekâ yazım imzalarını temizleyen Claude skill'i. YZ üretimi Türkçe'yi doğal, insan-sesli Türkçe'ye dönüştürür. | A Claude skill that removes AI writing signatures from Turkish text.
A sophisticated system for transforming AI-generated text into natural, human-like content using advanced NLP techniques and mathematical models.
Remove AI-like wording: style guidelines for Traditional Chinese writing and translation (Claude Code skill)
A powerful Model Context Protocol (MCP) server that helps refine AI-generated content to sound more natural and human-like. Built with advanced AI detection and text enhancement capabilities.
COLING'24 Humanizing Machine-Generated Content: Evading AI-Text Detection through Adversarial Attack
Find and repair substance defects in AI-assisted prose, code, docs, and agent output. Reports defects, never authorship. Structural tests over model judgement, because LLM judges agree with human slop labels at chance.
Elite frontend design taste for Claude Code / Cowork — a merged synthesis of emilkowalski/skill, pbakaus/impeccable, and leonxlnx/taste-skill. Typography, color, motion, components, anti-slop.
AI humanizer skill for Claude Code, Codex and Cursor. Removes AI slop and the signs of AI writing, then measures the rewrite against human baselines.
Agent skill that removes AI writing patterns from English, Russian and German prose. Merges stop-slop and avoid-ai-writing, adds a zero-dependency checker.
Sumi — UX/UI design intelligence for Claude Code: 43 skills, 168 references, 37 commands, an anti-slop engine, and an auditable citation trail. Start with /sumi:start.
Rewrite AI-generated text to sound human. 3-pass editing system with 36+ banned words, 10 structural patterns, and a quality checklist. Works with any LLM — Claude, ChatGPT, Gemini, Cursor, Windsurf. Also available as a Claude Code plugin.
Make AI-generated prose sound human. 29-pattern catalog + 8-rule operating system + 8-dimension scoring rubric. Works with Claude Code, OpenCode, Copilot CLI, Codex, Gemini, Goose.
[NeurIPS 2025] Implementation for paper "Adversarial Paraphrasing: A Universal Attack for Humanizing AI-Generated Text"
Humanizer Chinese academic version, Claude Code Skills/System Prompt, Chinese academic paper prompt for removing AI traces.
Open-source AI text detector and ChatGPT detector: 23 engines, one calibrated score, every number measured on a public benchmark. Self-hosted, runs on CPU.
ZAYUVALYA - AI Humanizer is a developing tool designed to rephrase AI-generated text to make it more natural and reduce detection by AI content detectors. Currently in its early stages, the project focuses on context-aware paraphrasing and aims to improve with further optimizations.
去除中文MBA论文AI写作痕迹的检测与改写工具 | AI writing pattern detection and removal tool for Chinese MBA thesis
PencilPlaybook is the UI Skills / Taste-Skill for Pencil.dev + Claude Code — a design playbook that gives Claude real perceptual psychology and senior-level guardrails so it stops producing averaged-out AI slop.
🚫 Stop AI slop in Chinese writing. 干掉中文 AI 味 — banned words, punctuation rules, structure constraints & 4-layer QA. Works with Claude Code, Cursor, Codex CLI.
HumanizerAI Agent Skills for Claude Code and Codex - AI detection and text humanization
A local model that humanize AI text and help you bypass every AI detector
Очеловечить текст нейросети: убирает длинное тире, канцелярит и штампы ChatGPT, не трогая факты. Правила для любой нейросети + скилл для Claude Code, Cursor, Codex. · Make AI text sound human: Russian + English rules and a Claude Code skill.
去 AI 味设计系统库 · 6 美学家族 × 64 套设计系统 · 复制一段提示词粘给任意 AI 就出片不飘 · 含在线预览站 | Anti-slop design systems for any AI — 64 curated design systems across 6 aesthetics, copy-paste prompts, live demo.
A Gemini CLI skill to make text more genuine, natural, and free of AI tropes.
50,000+ unique HTML presentation designs. Zero dependencies. Anti-AI-slop. A Claude Code skill.
this is the text-rewriter-python repository, an open-source project that provides a python script to "humanize" ai-generated text. the script makes text more natural and readable by performing operations such as sentence and word-level processing, part-of-speech tagging, synonym replacement, symbol preservation, and formatting improvements.
Ready-to-use design taste for Claude Code, Cowork, and Cursor. Prom = Thai for ready. Learn free at loop.sv-academy.org
Drop-in skills that stop AI coding agents from shipping garbage — kill-slop, security-review, PR hygiene
专门针对大学生学术写作（论文、实验报告）的去 AI 味与反检测提示词指南，将浮夸的机器腔调转化为朴实严谨的学生真实笔触。An anti-AI detection and text humanizer guide tailored for students' academic writing, converting robotic tones into authentic, rigorous student styles.
Content studio skill pack for coding agents — blog, social, slides, video, infographics — all brand-aware with built-in anti-slop. Works with Claude Code, Codex, and AGENTS.md.
中文 AI 文本去痕迹 + Claude 水印检查 — 六段式改写；零宽字符、同形字清干净，SynthID 采样水印只量残留、不声称能删。纯 Python 本地跑，零 LLM 零 API Key。Chinese AI-text humanizer & Claude / SynthID watermark inspector, fully offline, zero-dependency.
Anti-slop linter: catches comments that narrate change history instead of a current constraint (Python/JS/TS/Terraform)
Czech AI text humanizer (český humanizér AI textu). Detects and rewrites 27 AI writing patterns specific to Czech language. Dual-pass system: rewrite + self-check. 4 output styles. Works with Claude Code, ChatGPT, Gemini & any LLM.
Remove AI-like style from legal paper outlines: a Chinese legal writing skill based on real Law journal table-of-contents data.
An AI skill (Cursor/ Claude) for transforming AI-generated academic text into natural, human-like scholarly writing/ 一个专为社科学者设计的 Cursor AI Skills (Claude可用)，帮助将 AI 生成的英文学术文本转化为自然、人性化的学术写作。
A universal AI writing skill that turns generic drafts into specific, truthful, voice-consistent prose.
Create high-converting, visually distinctive landing pages with Vibe Discovery process and anti-AI-slop principles
Anti-slop content audit skill for X, ViewFT, LinkedIn, and long-form posts.
Strip AI writing tells at two layers: surface (Wikipedia Signs of AI writing) + narrative (StoryScope, arXiv:2604.03136). A model-agnostic LLM skill.
Two Agent Skills that help an AI agent make an interface clear, consistent, easy to use and scalable, and Russian text clear and correct (English version planned), without the signs of AI-generated work. They work in Claude, Codex, Cursor and other compatible agents.
An evidence-based AI writing skill for clearer, more human mathematical prose, built from stylometry and reference-style analysis
Agent Skill for editing prose so it does not read like generic LLM output
One-click removal of AI-like tone from WeChat, blog, and technical articles, plus distillation of your writing style using mainstream writing skills.
Natural Chinese and English writing with less AI-like style Skill: supports fact preservation, tone calibration, scenario routing, and auditable prompts.
Free anti-AI-slop system prompt & skill that makes ChatGPT, Claude, Gemini, Grok & Kimi write like a human. Bans AI tells (delve, tapestry, 'not just X but Y', em-dash overuse) and humanizes output.
An Agent Skill that makes AI-generated content sound less AI-generated, for direct generation and English translation.
DeepSeek Harness (DSH) novel plugin: code-enforced fact ledger, context pack, stage gates, free AI-style scan, deterministic audit, proposed revisions.
A Codex and Claude Code skill that rewrites AI-sounding Korean to sound natural and fit the context.
Korean humanizer prompt and skill for removing the usual AI smell from generated writing.
Claude Agent Skill that eliminates generic AI-generated design. Stop your AI shipping the same purple-gradient, Inter-font, three-feature-card website as everyone else.
Claude skill that goes past word-list humanizers: strips the full catalog of AI tells, refits the text to its channel (LinkedIn, X, cold email, Slack, exec memo, blog), and learns your voice from your own edits. Two-gate self-audit, strict no-fabrication rule.
Portable Agent Skill that removes AI writing patterns while preserving the writer voice. Works with Claude Code, Codex, OpenClaw, Cursor, and more.
A public writing skill for removing common AI patterns from prose.
Anti-AI-slop code analysis — detects machine-generated code patterns across Rust, TS/JS, Python, HTML/CSS
Anti-AI slop scientific writing skill for Claude Code. Or: Scientific writing with a human voice.
Claude Skill for removing Chinese AI writing: split three-part lists, use verbs, replace abstract subjects with details. Inspired by hardikpandya/stop-slop.
Fast Python linter for catching LLM-generated anti-patterns. Written in Rust, built on tree-sitter.
A deterministic Python linter that catches AI coding agents silencing type checkers without adding evidence.
🚫 Spot AI-generated YouTube scripts and content farms in real-time. Open-source browser extension for Chrome & Firefox with thumbnail badges and sentence heatmaps.
Don't be a sloperator. Ten rules that stop AI from writing and behaving like naked AI. Principle-based, not vocabulary bans. Includes a bonus /work skill that forces thinking before output. Works with Claude Code, ChatGPT, Gemini, or any LLM.
Free Claude skills for SEO content: AI humanizer, AI detection bypass, keyword preservation, agency QC, local SEO, programmatic SEO, content repurposing. Drop-in markdown for Claude projects. MIT licensed.
Writing skill that removes AI-like style, with a 5-part scoring system detecting 41 AI patterns. Makes Claude/Cursor/Copilot write naturally. Claude Code Skill…
Anti-AI-slop agent skill: makes text read written and UI look made, not generated. Deterministic scanners + severity-tiered tell catalogs.
Skills for AI agents to remove or prevent slop in generated frontend UI design and written content.
Reduce the "AI-generated" feel in business and academic texts; applicable to both Chinese and English. 降低商务与学术文本的 AI 生成感，中英文均适用。
Claude Code skill for detecting AI-generated writing patterns (slop) in content
Claude Code skill (/humanize). Strips AI writing patterns from content via five editorial layers: vocabulary, sentence shape, bullet rhythm, paragraph structure, and section narration.
Detect 47 AI writing patterns and rewrite text in 5 human voice profiles. A Claude Code skill backed by ICML 2025 research on token-probability distributions.
Skill for lowering AI detection scores on CNKI, VIP, Turnitin, and more. Supports Chinese and English; keeps language clear and academic. https://ushallpass.ai/
AI text humanizer skills for Arabic (MSA, Egyptian, Levantine) and Hebrew — zero AI tells
AI Humanizer is a free, open-source tool that rewrites AI-generated text to sound more natural, fluent, and human.
Make AI-generated text sound more natural in four styles: professional business, technical, conversational, and academic.
A customizable, maintainable agentic system for writing, editing, rewriting, polishing, and reviewing Chinese text without an AI-generated feel.
Chinese fiction de-AI writing skill for AI-assisted novel creation
AI humanizer: CLI tool and Claude Code skill for rewriting AI-generated text into natural human writing.
Claude Code skill для чистки русских текстов от AI-штампов. Адаптация blader/humanizer: канцелярит, кальки, раздутая значимость + типографическая деградация и режим аудита.
Prose linter + AI-slop detector: weasel words, passive voice, hedging, and 190+ research-cited AI tells. Web editor, API, MCP server, CLI, GitHub Action.
Fork français du skill Claude Code « humanizer » (blader/humanizer, MIT) — 33 motifs d'écriture IA adaptés au français
Traditional Chinese blader/humanizer: humanizes AI Chinese with 49 patterns for China terms and punctuation; supports Claude Code, Codex, and other agents.
Humanizer is an AI writing skill that detects and rewrites common signs of AI-generated prose. It helps agents preserve meaning, match audience and tone, reduce hype and filler, and produce writing that sounds natural, specific, and human across drafts, posts, docs, and messages.
Dwujęzyczny (PL+EN) humanizer / de-slop skill dla Claude Code. Fork blader/humanizer + warstwa polskiej naturalności + eksperyment: czy Bielik (polski LLM) jest dobrym recenzentem polszczyzny.
Claude skill that removes AI writing patterns from Swedish professional text. Covers 16 pattern types including significance inflation, nominalization overuse, passive voice as false formality, anglification, and bullet/emoji formatting. Tuned for Nordic professional voice.
繁體中文（臺灣）AI 味清理 Skill：去 AI 腔、清中港用語、修簡轉繁錯字、教育部標點、七種出版語體。AI humanizer skill for Traditional Chinese (Taiwan) — Claude Code / Codex / Cursor / ChatGPT.
An Agent Skill that makes LLM-written Korean sound natural to Korean readers while preserving facts, numbers, and domain-specific terminology.
Taste Skill Simplified Chinese edition - anti-slop frontend design skills for AI agents
Production-grade, anti-slop AI writing suite for agent harnesses with voice calibration, craft rules, de-slop editor, and deterministic AST linting
A shot-by-shot review and refinement skill for removing the AI feel from Chinese web novels.
Remove AI-like writing from text, papers, and paragraphs: use AI to lower AI detection scores, using magic to defeat magic.
🎨 Create distinctive static sites with React, Tailwind, shadcn/ui — no mockups needed. Anti-AI-slop design, mobile-first patterns, single-file bundling.
Claude Code skill for removing signs of LLM-generated papers. Covers 28 patterns with four-layer checks, from Zhihu (up to 757 likes) and Wikipedia.
繁體中文寫作 skill 工具鏈 for Claude Code：humanizer-tw（去 AI 味）+ good-writing-zh（節奏打磨：PG／余光中／王鼎鈞 + 技術文件保守模式）+ fable-econ／fable-explore（Amanda Askell 寓言式概念學習）
A Claude Agent Skill that humanizes AI-written marketing and long-form copy: landing pages, sales pages, e-books, case studies, founder notes. Removes AI tells and process bleed, de-slops prose, strips em dashes, never invents facts, keeps your brand voice.
Write like some of the best writers of our time. This skill goes beyond removing AI patterns from writing and writes like an actual human
Agent-agnostic skill for writing natural, humanized, SEO-focused articles and blog content without robotic AI patterns.
Claude Code plugins from Floka AS — Norwegian-flavored AI tooling: humanizer (AI-writing scrub + Norwegian patterns), board (build a personal advisory board), translate-to-no (English → idiomatic Norwegian), forge (our personal coding harness for contracts and automated development)
Opinionated Biome GritQL rules for rejecting low-evidence TypeScript patterns
Tool for making Chinese text sound less AI-generated. 34 Patterns (29 original + 5 Chinese-specific) for Chinese writing.
Free Claude Code skill to humanize AI-generated text. Bypass GPTZero, Turnitin, and other AI detectors. Powered by DiaIQ.
Monorepo workshop for agent-development tooling: orchestrator-workflow and okf-kit ship on npm; slop-detector is the AI-slop linter for PRs.
Evidence-first, anti-slop opportunity discovery skill for coding agents (Claude Code + Codex)
Three small, independent guardrails for AI coding agents: an anti-AI-slop writing checklist, a code-comment hygiene checklist, and a standing multi-agent orchestration recipe hook.
Measure and stop structural erosion in AI-written TypeScript. Slop Check finds the mess, the Deslop tools clean it. Based on SlopCodeBench (arXiv:2603.24755).
A Claude skill that strips AI writing patterns and rewrites text to sound human. Synthesized from 8 humanizer repos.
Humanize AI-written text — a portable agent skill that writes or rewrites prose to remove AI tells. Works in Claude Code, Cursor, Codex, OpenClaw.
Free content skills for AI agents. Humanize AI text, keep your keywords intact, ship agency-quality drafts. Free, MIT, works across Claude Code, Codex, Cursor and ChatGPT.
Claude Code skill that removes signs of AI-generated writing from Japanese text
A Claude Code skill that strips AI tells out of cold email. Anti-slop engine for startup outbound to technical people.
Turkish humanizer skill for Claude and Claude Code. Rewrites AI-sounding Turkish into Turkish a person would have written: native sentence architecture, register-aware dosing, technical terms kept. Türkçe humanizer / yapay zeka metni insanlaştırma.
Magyar írásstílus-skill agenteknek: a kimenet ne hangozzon AI-nak. Írás közben ad szabályokat, nem a kész szöveget javítja utólag. Csak azt állítja, amit a forrás tartalmaz. 26 minta, dokumentumszintű döntések és LLM nélkül futó lint. Claude Code, Codex, Cursor, Copilot, Gemini CLI és 74 további agent. Hungarian humanizer / anti-AI-slop skill.
Chinese AI-style writing cleanup skill: uses density thresholds and a reverse checklist instead of the popular AI-style feature checklist.
A Chinese academic writing editor for Codex that reviews and rewrites papers and research proposals to remove templated AI-style wording.
Jev 辅助去 AI 味写作：保留事实、归因与作者语气 | Natural prose editing with Jev-assisted fidelity review
A Claude skill that answers exam questions exactly like a university student — concise, humanized, and always referenced to Iranian university textbooks. No extra explanation, just the right answer.
Mr.Li Writer is an anti-AI-tone Chinese long-form writing skill. It first evaluates whether your idea is worth writing and helps you find a distinctive angle, then designs a unique opening and title for each piece, and delivers the final draft in a concrete, opinionated, humanized voice—ready to publish across platforms.
library of ui references that prevent ai's from creating vibecode slop
多语言 AI 味检测与去 AI 化工具集 · Multilingual AI-slop detector & de-slopping toolkit (zh/en/es/ar/hi)
A Python package for humanizing AI-generated or robotic text using techniques like DIPPER, HMGC, and ADAT.
CLI tool to detect AI-generated text patterns and improve writing to avoid detection. Score, analyze, and humanize content.
人味 Renwei — 书稿级中文去 AI 味 / Book-length Chinese AI humanizer. 51 rules covering AI tells, Euro-Chinese syntax, whole-book consistency & positive prose craft. Ships a token-free lint CLI. AGENTS.md compatible.
Chinese LLM writing Skill for reducing predictable AI prose patterns without changing facts or author voice.
15 free art-directed website build-prompts + an anti-slop design SKILL.md for Lovable, Bolt, v0, Cursor, and Claude Code — from Meez (meez.design)
Anti-slop frontend design skill (universal frontend design + deck mode) with 30 real brand material libraries, manifest-driven style selection, and QA gates
| Tool | Stars | Score |
|---|---|---|
| taste-skill | ★ 93.1k | 85 |
| humanizer | ★ 52.7k | 84 |
| Humanizer-zh | ★ 19.1k | 77 |
| anti-ai-slop-writing | ★ 323 | 65 |
| humanizer_academic | ★ 271 | 75 |
| hallmark | ★ 28.7k | 83 |
| stop-slop | ★ 17.8k | 64 |
| im-not-ai | ★ 5.7k | 74 |
| avoid-ai-writing | ★ 4.9k | 77 |
| anti-slop | ★ 4.6k | 84 |
| anti-slop | ★ 4.3k | 82 |
| desloppify | ★ 3.2k | 60 |
| humanize-text | ★ 3.0k | 71 |
| sepia | ★ 3.0k | 75 |
| lieflat-less-ai-tone | ★ 2.4k | 63 |
| AIWriteX | ★ 2.0k | 66 |
| shuorenhua | ★ 2.0k | 69 |
| kill-ai-slop | ★ 1.2k | 76 |
| speak-human-tw | ★ 1.0k | 70 |
| auteur | ★ 1.0k | 75 |
blader/humanizer is an agent skill that finds and removes signs of AI-generated writing from text; op7418/Humanizer-zh is its Chinese version. Related skills apply the same idea elsewhere: hardikpandya/stop-slop for prose rules, Leonxlnx/taste-skill and Nutlope/hallmark for UI design, peteromallet/desloppify for code.
Install the skill folder (it contains a SKILL.md) into the agent's skills directory, then ask the agent to humanize or de-slop a piece of text; the skill supplies the patterns to look for and the rewrite rules. Every humanizer on this page is in the open-source list zhuyansen/awesome-humanizer-skills on GitHub, with its security grade.
The top humanizer & anti-slop skills in 2026 are taste-skill, humanizer, Humanizer-zh. Agent Skills Hub ranks 229 options by GitHub stars, quality score (6 dimensions including completeness, examples, and agent readiness), and recent activity. The list is rebuilt every 8 hours from live GitHub data.
taste-skill (93.1k stars) is the most adopted choice for general humanizer & anti-slop skills workflows, written in JavaScript. humanizer (52.7k stars) is a strong alternative and uses Python instead. Pick by your existing stack: match the language and runtime your team already uses to minimize integration cost. If unsure, start with taste-skill — it has the deepest community and the most examples online.
Avoid pre-built humanizer & anti-slop skills when (1) your use case requires deep customization that the tool's plugin system doesn't support, (2) you have strict compliance requirements that ban third-party dependencies, (3) the tool's maintenance is inactive (last commit >6 months ago), or (4) your data volume is small enough that a 50-line custom script is cheaper than learning the tool. For most production workflows above 100 requests/day, the time savings from a maintained tool outweigh the customization loss.
Humanizer & Anti-Slop Skills focuses specifically on open-source humanizer and anti-slop skills, security-graded: make ai-written text, code and ui read like a person made it (blader/humanizer, humanizer-zh, taste-skill), and detect ai tells. vet before you install. AI Design Tools is a related but distinct category — see https://agentskillshub.top/best/ai-design/ for those tools. The two often appear in the same agent pipeline but solve different problems: choose humanizer & anti-slop skills when your primary goal is the specific task, and ai design tools when the workflow is broader.
For most teams, yes. taste-skill has 93.1k stars worth of community testing, handles edge cases you haven't thought of, and ships with documentation. Build your own only when (1) your requirements are deeply non-standard, (2) you have a security/compliance reason to avoid OSS dependencies, or (3) the maintenance burden is small enough (<200 lines of code) that you'll save time long-term. The break-even point is usually around 2-3 weeks of dev time saved.
Most humanizer & anti-slop skills listed are open source under permissive licenses (MIT, Apache 2.0). A handful offer paid managed/cloud versions on top of free self-hosted core. Always check the LICENSE file on each tool's GitHub repository before commercial use — some use AGPL or non-commercial restrictions that may not fit your deployment model.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: