175,000+ GitHub Stars! Top 10 Chinese AI Creators' Open-Source Agent Skills Portfolio (Visual Samples & One-Click Install)
2026-10-10 · bestaiskill.com
175,000+ GitHub Stars! Top 10 Chinese AI Creators' Open-Source Agent Skills Portfolio (Visual Samples & One-Click Install)
Introduction: Why Do Your AI Outputs Feel "Synthetic," While Top Creators Ship Viral Hits?
Many users run into the same frustrations: "Why do my AI-generated articles sound like robotic machine translations? Why do AI illustrations look like generic, tacky stock photos? Why do AI-generated pitch decks still look like cheap templates?"
The answer is simple: **You are still prompt engineering in raw conversational chats, while top-tier creators have already engineered their battle-tested workflows into production-grade Agent Skills.**
An Agent Skill is far more than a "prompt template." It is a full-fledged productivity suite comprising **deterministic workflows, design systems built with high aesthetic taste, strict code-level validations, and reusable asset templates.**
Across the open-source projects of the 10 leading creators highlighted here, their repositories have garnered **over 175,200 Stars and 18,100+ Forks on GitHub**, with flagship projects surpassing **34,000 Stars** individually.
To help developers and power users evaluate their impact and securely integrate them into production agents, **skills.bayx.uk (SkillPub Teardown Lab)** conducted **line-by-line static source code audits, locked commit snapshot archives, and set up high-speed distribution mirrors** for every repository featured below.
Here is an architectural breakdown of their signature tools, accompanied by **visual output comparisons, hard open-source data, and one-click copyable installation commands**.
---
1. Baoyu (JimLiu) — Industrialized Digital Publishing Pipeline
**GitHub Metrics**: ⭐ **26,500+ Stars** | 🍴 **2,900+ Forks** (Repository: `JimLiu/baoyu-skills`)
**Creator Profile**: Renowned software architect, technical translator, and prolific AI creator. Baoyu pioneered the systematic industrialization of technical content publishing, emphasizing rigorous multi-stage interactions and clear output schemas.
**Core Use Cases**: Long-form editorial visual placement, educational multi-panel comics, high-fidelity slide decks, and automated cross-platform distribution.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: Moving beyond "random AI images tossed into an article," Baoyu's `baoyu-article-illustrator` employs a 3D orthogonal matrix of "Content Type × Artistic Style × Color Palette," placing visuals strictly where analytical arguments require them. Meanwhile, `baoyu-comic` preserves visual continuity across panels to generate publication-grade educational comic strips.
Featured Skills:
**baoyu-article-illustrator (Editorial Illustration Expert)**: Analyzes article structure and information density, matching visual schemas via a "Type × Style" orthogonal matrix to place purposeful illustrations strictly at argument anchors.
baoyu-article-illustratorL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandWrite FileJimLiu/baoyu-skills
Analyzes article structure, identifies positions requiring visual aids, generates illustrations with Type × Style × Palette three-dimension
**baoyu-comic (Educational Comic Studio)**: Deconstructs complex conceptual ideas into four-panel or multi-panel narrative comics, automatically enforcing character consistency and exporting locally to PDF or web visuals.
baoyu-comicL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandWrite FileJimLiu/baoyu-skills
Knowledge comic creator supporting multiple art styles and tones. Creates original educational comics with detailed panel layouts and batch-
**baoyu-slide-deck (Presentation Deck Architect)**: Transforms structured outlines into 16:9 widescreen presentation card decks, balancing visual rhythm with readability.
baoyu-slide-deckL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork AccessJimLiu/baoyu-skills
Generates professional slide deck images from content. Creates outlines with style instructions, then generates individual slide images. Use
---
2. Li Jigang (lijigang) — Cognitive Anatomy, Feynman Distillation & Philosophical Card Crafter
**GitHub Metrics**: ⭐ **7,480+ Stars** | 🍴 **840+ Forks** (Repository: `lijigang/ljg-skills`)
**Creator Profile**: Went viral across the tech ecosystem with his philosophical "Lisp-style cognitive cards." Li excels at deconstructing complex human cognition, mental models, and classic theories into computational schemas, characterized by a sharp minimalist aesthetic.
**Core Use Cases**: Conceptual deconstruction, intuitive Feynman-style explanations, cognitive blind-spot diagnosis, business model canvases, and 1080px infographic card generation.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: The viral "New Chinese Dictionary" format originated here. Powered by a Bun + Playwright rendering pipeline, it applies strict typography principles—deliberate whitespace, golden ratios, and subtle skeuomorphic borders—to forge several hundred words of insight into a meticulously typeset, highly shareable 1080px image card.
Featured Skills:
**ljg-card (Cognitive Card Crafter)**: The iconic Lisp conceptual card system, leveraging a headless Bun and Playwright runtime to transform core insights into elegant visual cards, comic layouts, or whiteboard sketches.
ljg-cardL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accesslijigang/ljg-skills
"Content caster (铸). Transforms text into PNG through precise HTML typography and, when the mold needs it, generated raster imagery. Four mo
**ljg-explain (Feynman Explainer)**: Explains abstract concepts, mathematical mechanics, and underlying logic in plain terms until users can independently derive the ideas and apply them across domains.
ljg-explainL1 Basic · Source registration onlyV1 Commit-Pinnedlijigang/ljg-skills
"Explains concepts, formulas, mechanisms, and principles until they can be recognized, derived, and transferred: connecting specific scenari
**ljg-blind (Cognitive Blind Spot Scanner)**: Analyzes conversation histories and reasoning traces to uncover implicit assumptions and self-imposed constraints, recommending targeted reading chapters to close the loop.
ljg-blindL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accesslijigang/ljg-skills
Blind spot scanning—reads all your dialogues with AI from yesterday to illuminate exposed cognitive blind spots (not lacking knowledge, but
---
3. Guizang (op7418) — High-Aesthetic HTML Decks & Natural Language Humanizer
**GitHub Metrics**: ⭐ **47,000+ Total Stars** (`guizang-ppt-skill`: 27.6k Stars + `Humanizer-zh`: 19.4k Stars)
**Creator Profile**: Digital nomad, senior designer, and AI practitioner. Guizang is dedicated to pristine design standards and natural prose, making his tools favorites among discerning designers and engineers.
**Core Use Cases**: Native HTML/CSS/JS interactive presentations, authentic natural-language text polishing, and social media visual assets.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: Moves completely away from rigid, legacy presentation software. `guizang-ppt` draws inspiration from the Swiss International Typographic Style and modern frontend web transitions to compile standalone, responsive HTML presentation files with keyboard navigation. Complementing this, `humanizer-zh` eliminates robotic AI clichés and bloated corporate jargon, trained across tens of thousands of refined text samples.
Featured Skills:
**guizang-ppt (High-Fidelity HTML Presentations)**: Replaces cookie-cutter presentation decks with standalone web-first applications supporting full-screen interactivity, theme switching, and fluid responsive layouts.
guizang-pptL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accessop7418/guizang-ppt-skill
Generates horizontal swipe web-based presentation decks (single HTML file), featuring WebGL backgrounds, presenter view, audience screen syn
**humanizer-zh (Natural Language Polisher)**: Systematically removes synthetic AI cadence, clunky translation tones, and hollow filler phrases, restoring a direct, punchy, and natural voice.
humanizer-zhL2 Local · Restricted workspaceV1 Commit-PinnedWrite Fileop7418/Humanizer-zh
Removes traces of AI-generated text. Ideal for editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia
---
4. Xiaohu (xiaohuailabs) — Original IP Asset Library & Contextual Illustration Studio
**GitHub Metrics**: ⭐ **300+ Early Pioneer Stars** (Repository: `xiaohuailabs/xiaohu-ip-studio`)
**Creator Profile**: Author of the "Xiaohu" technical publication, focused on practical AI deployment. He addresses the issue of disjointed, generic technical illustrations by introducing a recurring "Character IP" visual framework.
**Core Use Cases**: Illustrated long-form technical guides, continuous brand character identities, and witty visual metaphors.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: Avoids disjointed, mishmashed imagery. Xiaohu open-sourced an asset library of **31 original IP characters** (including 15 hand-drawn line-art figures and 16 meme mascots). Guided by a three-step pipeline—"Identify Anchor → Script Metaphor → Anti-Template Audit"—a consistent character guides readers through the entire article, building immediate visual recognition.
Featured Skills:
**xiaohu-ip-studio (IP Illustration Studio)**: Integrates 31 original IP characters (15 line-art figures + 16 expressive meme characters). Uses an structured workflow to cast custom characters into continuous roles across long-form technical essays.
xiaohu-ip-studioL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accessxiaohuailabs/xiaohu-ip-studio
Open-source Chinese in-text illustration skill + IP character library. Using the methodology of "identify cognitive anchors → construct meta
---
5. Digital Life Khazix (KKKKhazix) — High-Engagement Deep-Dive Long-Form Writing
**GitHub Metrics**: ⭐ **21,300+ Stars** | 🍴 **2,250+ Forks** (Repository: `KKKKhazix/khazix-skills`)
**Creator Profile**: Leading AI business analyst and writer, celebrated for distinctive first-person long-form technical deep dives and market analyses that consistently attract hundreds of thousands of readers.
**Core Use Cases**: High-impact editorial writing, multi-source global AI news synthesis, and sector-wide technical teardowns.
Core Strengths
> **The Writing Edge**: Most AI-written articles read like dull technical manuals that lose readers within paragraphs. Khazix's writing skill models his signature structure: emotional progression, accessible analogies, and immersive first-person deductions. The resulting long-form content maintains dense technical signal while remaining compelling and human.
Featured Skills:
**khazix-writer (Signature Style Writing Copilot)**: Models an engaging narrative methodology that balances conceptual analogies with layered analytical deductions, helping authors craft high-retention long-form publications.
khazix-writerL1 Basic · Source registration onlyV1 Commit-PinnedKKKKhazix/khazix-skills
Long-form WeChat Official Account writing skill modeled after Digital Life Khazix. Use this when users need to draft WeChat articles, write
**aihot (AI Intelligence Aggregator)**: Ingests and synthesizes major global AI developments from diverse feeds, quickly isolating commercially viable technological breakthroughs from market noise.
aihotL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork AccessKKKKhazix/khazix-skills
Query AIHOT for Chinese AI news, curated picks, current trending topics, and daily briefings. Used when users ask about today's or recent AI
---
6. Huashu (alchaincyf) — Frontend Hacker Design System & Nuwa Thought Architect
**GitHub Metrics**: ⭐ **58,800+ Total Stars** (`nuwa-skill`: 34k Stars + `huashu-design`: 24.8k Stars)
**Creator Profile**: Senior frontend engineer and interaction designer. Huashu bridges production code with design theory, enforcing strict "Anti-Slop" standards alongside distilled mental models of renowned leaders.
**Core Use Cases**: High-aesthetic frontend interactive prototypes, motion-design exports, and strategic simulations modeled after 15 world-class entrepreneurs.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: Enforces a strict "Three-Direction Gate." For any interface or component request, the skill generates 3 divergent design explorations upfront, followed by a mandatory "Anti-Slop Checklist" to eliminate generic enterprise templates.
Featured Skills:
**huashu-design (Full-Stack Design Studio)**: End-to-end native HTML/CSS/JS design system. Enforces the "Three-Direction Gate," strict anti-slop verification checklists, and programmatic video transition exports.
huashu-designL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accessalchaincyf/huashu-design
Huashu Design — Create high-fidelity prototypes, presentation decks, animations, data visualizations, and expert design reviews using HTML.
**steve-jobs-perspective (Steve Jobs Strategic Advisor)**: Distilled from the Nuwa persona library. Drawing from decades of keynotes, interviews, and strategic inflection points, it evaluates product roadmaps through Steve Jobs' product intuition.
steve-jobs-perspectiveL1 Basic · Source registration onlyV1 Commit-Pinnedalchaincyf/nuwa-skill
Steve Jobs' thinking framework and communication style. Based on deep research across Walter Isaacson's authorized biography, the Stanford c
---
7. Xiangyang Qiaomu (joeseesun) — Mondo-Style Movie Posters & Automated Video Pipelines
**GitHub Metrics**: ⭐ **500+ Stars** | Cinematic visual systems and automated video production pipelines (`joeseesun/qiaomu-*`)
**Creator Profile**: Veteran automation engineer and knowledge architect. Qiaomu connects fragmented generative media tools into unified, industrial workflows.
**Core Use Cases**: Single-prompt video project synthesis, Mondo-grade poster generation, and multimodal knowledge base ingestion for NotebookLM.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: Distills the visual signatures of 20 world-class Mondo movie poster artists (such as Olly Moss and Tyler Stout) into generation presets. A single prompt yields 4 distinct artistic interpretations; simultaneously, his automated director workflow translates narrative ideas directly into renderable video timelines.
Featured Skills:
**qiaomu-codex-imagegen (Mondo Visual Poster Workshop)**: Embeds visual presets inspired by 20 Mondo poster masters, generating 4 aesthetically divergent compositions prior to final image rendering.
qiaomu-codex-imagegenL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accessjoeseesun/qiaomu-codex-imagegen
Generate or edit images with Codex's built-in image generation from any agent, through an MCP tool or a plain CLI. For every image request i
**qiaomu-cut (Intelligent Video Director)**: Converts a single prompt into an end-to-end, reproducible video project timeline, providing 12 pre-configured pipelines for code tutorials, explainer reels, and more.
qiaomu-cutL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accessjoeseesun/qiaomu-cut-skill
An intelligent editing director, Qiaocut, that transforms one-line requests into reproducible and verifiable video projects. Use when the us
---
8. Huangshu (Backtthefuture) — Executive Advisory Board & Practical Workflow Engines
**GitHub Metrics**: ⭐ **270+ Practitioner Stars** (Repository: `Backtthefuture/huangshu`)
**Creator Profile**: Seasoned tech veteran and business operator, focusing on high-frequency commercial decision-making, executive recruiting, and multimodal asset extraction.
**Core Use Cases**: 12-expert executive advisory roundtables, verbatim cross-platform audiovisual transcription, and interactive hiring workflows.
Core Strengths
> **The Strategic Edge**: Breaks beyond basic single-turn question-and-answer prompts. `advisory-board` assembles a virtual board of 12 iconic strategists—including Warren Buffett, Charlie Munger, Jeff Bezos, and Steve Jobs—driving a structured sequence of cross-examination, defense, and synthesis around strategic business decisions.
Featured Skills:
**advisory-board (Virtual Executive Board)**: Brings together 12 influential thinkers (Buffett, Munger, Jobs, and others) across structured boardroom stages to pressure-test critical strategic decisions.
advisory-boardL2 Local · Restricted workspaceV1 Commit-PinnedWrite FileBacktthefuture/huangshu
Uncle Huang's Advisory Board — A business decision think tank composed of 12 top thinkers. Uses a structured advisory board process to let d
**video-transcript (Verbatim Multi-Platform Transcriber)**: Extracts and transcribes audio from YouTube, TikTok, Bilibili, and other platforms into timestamped, verbatim records that capture natural speech disfluencies.
video-transcriptL5 Restricted · Isolated sandbox & strict auditV2 Human-AuditedTerms / ComplianceExecute CommandBacktthefuture/huangshu
Video verbatim transcript extraction expert (powered by Doubao video understanding model). Supports links from Bilibili, Douyin, Xiaohongshu
---
9. Canghe (freestylefly) — End-to-End Media Automation & Omnichannel Publishing
**GitHub Metrics**: ⭐ **460+ Developer Stars** | 🍴 **100+ Forks** (Repository: `freestylefly/canghe-skills`)
**Creator Profile**: Technical evangelist and content automation engineer. Canghe extends community tooling into production-ready pipelines that automate the entire path from ideation to distribution.
**Core Use Cases**: Contextual illustration layouts, episodic video script generation, commercial intelligence dashboards, and one-click publishing across WeChat and X (Twitter).
Core Strengths
> **The Automation Edge**: Bridges the gap between content generation and real-world distribution. From generating structured visual assets to driving Chrome via the Chrome DevTools Protocol (CDP) to inject inline styles into platform drafts, this skill frees creators from repetitive manual formatting.
Featured Skills:
**canghe-article-illustrator (Contextual Visual Architect)**: Generates structured prompts across multi-dimensional criteria, matching aspect ratios and stylistic needs across publication layouts.
canghe-article-illustratorL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandWrite Filefreestylefly/canghe-skills
Analyzes article structure, identifies positions requiring visual aids, generates illustrations with Type × Style two-dimension approach. Us
**canghe-post-to-wechat (WeChat Draft Automation)**: Handles the final step of converting Markdown into platform-native inline HTML, driving browser automation via CDP to populate drafts automatically.
canghe-post-to-wechatL5 Restricted · Isolated sandbox & strict auditV2 Human-AuditedTerms / ComplianceCredential / Sessionfreestylefly/canghe-skills
Posts content to WeChat Official Account (微信公众号) via API or Chrome CDP. Supports article posting (文章) with HTML, markdown, or plain text inp
---
10. Ian / Xiaohei (helloianneo) — Minimalist Hand-Drawn Editorial Illustrations
**GitHub Metrics**: ⭐ **12,500+ Stars** | 🍴 **1,400+ Forks** (Repository: `helloianneo/ian-xiaohei-illustrations`)
**Creator Profile**: Senior editorial designer and illustrator who established the recognizable "Xiaohei monochrome line-art" style for long-form essays.
**Core Use Cases**: Reflective, understated hand-drawn line illustrations designed to explain technical topics while eliminating synthetic visual artifacts.
Visual Breakthroughs & Practical Highlights
> **The Aesthetic Edge**: Crisp white backgrounds, minimalist black ink linework, and generous negative space. By avoiding loud gradients and hyper-rendered 3D textures, this style pairs cleanly with dense technical essays, giving editorial layouts breathing room and visual weight.
Featured Skills:
**ian-xiaohei-illustrations (Minimalist Line Art Studio)**: Built for deep technical and business essays, translating core thesis points into understated black-and-white line art that integrates smoothly with long-form text.
ian-xiaohei-illustrationsL4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandNetwork Accesshelloianneo/ian-xiaohei-illustrations
Generates Ian-style body illustrations for Chinese text. Designed for requests to create "grotesque/quirky," "Xiaohei (Little Black)," "hand
---
How to Install in Your Agent
No complex terminal commands or manual environment setups required. Every skill card includes a direct installation path:
1. **Copy Command**: Copy the installation command directly from the gray code block beneath the respective Skill card. For example:
`Install "ljg-card" using this manifest guide: https://skills.bayx.uk/skills/ljg-card.install.json`
2. **Send to Your Agent**: Paste the text into your AI coding assistant or agent platform (such as Claude Code, OMP, Cursor, or Codex). The agent will parse the manifest, pull verified snapshot files from the mirror, and finalize the setup automatically.
Every Skill details page also publishes its full **two-pass source code security audit report (covering operational logic, network egress restrictions, and audited commits)** to support safe deployment in production environments.