This project provides a specialized AI "Codex Skill" for generating unique, hand-drawn illustrations for Chinese articles, transforming abstract concepts into memorable visual metaphors. It targets content creators, knowledge workers, and AI users seeking a distinctive, quirky, yet clean visual style that goes beyond generic infographics or stock imagery, enhancing article comprehension and brand recognition.

How It Works

The skill operates by first analyzing Chinese articles, blog posts, or Notion content to identify key cognitive anchors—such as judgments, processes, states, or metaphors. It then guides an AI agent to conceptualize and generate a 16:9 horizontal illustration based on one of these anchors. The core approach emphasizes creating a "quirky but clean" aesthetic using a minimalist "Xiaohei" IP (a simple black character) actively participating in the depicted action, set against a pure white background with sparse, hand-written Chinese annotations in red, orange, and blue. This method prioritizes translating abstract ideas into concrete, memorable visual explanations.

Quick Start & Requirements

• Installation: Clone the repository and copy the ian-xiaohei-illustrations/ subdirectory into your Codex skills directory (e.g., ${CODEX_HOME:-$HOME/.codex}/skills/).
• Prerequisites: Requires a Codex environment. No specific hardware (GPU/CUDA) or Python version is explicitly listed.
• Documentation: Examples and usage prompts are available in examples/prompts.md.

Highlighted Details

• Generates 16:9 horizontal illustrations with a pure white background, fine black hand-drawn lines, and minimal red, orange, and blue Chinese annotations.
• Each illustration focuses on a single core concept, with the "Xiaohei" IP actively performing the central action.
• The visual style is intentionally "quirky, creative, clean," avoiding overly complex or childish aesthetics.
• Outputs a detailed "shot list" (theme, core meaning, structure type, Xiaohei action, annotation suggestions) prior to image generation.

Maintenance & Community

The project is maintained by Ian (伊恩), a product designer and AI builder. Related projects include "Ian Handdrawn PPT" and "Awesome Claude Code Skills." Community engagement and further exploration are encouraged through "AI 开眼日报" and the "Indie Builders Club." Contact information includes GitHub, X/Twitter (@ianneo_ai), and WeChat (ianneoxyz).

Licensing & Compatibility

The project is released under the MIT License, which permits broad use, including commercial applications and derivative works, provided attribution is maintained.

Limitations & Caveats

Generated images may exhibit AI-specific artifacts like typos, hallucinated labels, or style drift, requiring post-generation checks. The stability of Chinese text annotations is higher with shorter phrases. Illustrations are designed to convey a single core concept and are not intended as comprehensive infographics or editable vector files. The "Xiaohei" character must be integral to the action; if the image functions without him, the design may be too decorative. Example images serve as style guides, not composition templates. This skill is unsuitable for commercial illustration, branding KVs, or traditional presentation graphics.