A Codex Skill that guides AI agents to generate white-background, hand-drawn 16:9 illustrations for Chinese articles, featuring the Xiaohei character.
12.7K1.5KMITUpdated 27 days ago
GiMi-Xiaomi/gimi-illustration-skill
An AI Agent Skill that turns Chinese article or spoken-script text into readable illustrations, with custom characters, three styles, and adjustable size and titles.
- Stars
- 711
- Forks
- 78
- Open issues
- 0
- Last commit
- 3 months ago
What is it?
gimi-illustration is an open-source AI Agent Skill for generating illustrations from article or script text. It aims to turn judgments, workflows, comparisons, states and metaphors into clean diagrams that a reader can understand at a glance, and it is aimed at long-form articles, social image posts, blogs, Notion pages and spoken scripts.
The skill offers three visual styles: a default quirky hand-drawn line style, a warm watercolor storybook style, and a flat editorial style for product proposals and decision explanations. Users switch styles with natural-language requests rather than fixed IDs. Images can be made in landscape 16:9 by default or in portrait for formats such as Xiaohongshu, and an in-image title can be added on request.
Characters are pluggable. The default presenter is the Gimi character, and users can record their own character from a front-view image, or turn characters off entirely. Generation relies on the built-in image tools of the host platform, such as Codex or Cursor, and outputs are saved to an outputs folder along with a shot configuration file.
Key features
- 01Three switchable visual styles: quirky hand-drawn, warm storybook, and flat product proposal.
- 02Custom character support by recording a front-view image of your own persona.
- 03Characters can be turned off for diagram-only output.
- 04Landscape 16:9 by default, with portrait mode on request.
- 05Optional in-image titles specified in the prompt.
- 06Saves a shot-config file alongside outputs for later review.
Use cases
- Adding illustrations to long-form Chinese articles or blog posts.
- Creating vertical image cards for Xiaohongshu-style social posts.
- Illustrating methodology, workflow or decision explanations for internal proposals.
- Producing storyboard-style visuals for spoken video scripts.
Quick start
- 1
Clone the repository
git clone https://github.com/GiMi-Xiaomi/gimi-illustration-skill.git - 2
Enter the project folder
cd gimi-illustration-skill - 3
Copy the skill into the Codex skills folder
mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills" && cp -R . "${CODEX_HOME:-$HOME/.codex}/skills/gimi-illustration" - 4
In your chat, send 配图 followed by your article text
Things to know
- Image generation depends on the built-in tools of the host platform (Codex Image Gen or Cursor GenerateImage).
- The Gimi character artwork is not covered by the MIT license and needs separate permission to reuse.
Topics
FAQ
- What is gimi-illustration-skill?
- An AI Agent Skill that turns Chinese article or spoken-script text into readable illustrations, with custom characters, three styles, and adjustable size and titles. gimi-illustration is an open-source AI Agent Skill for generating illustrations from article or script text. It aims to turn judgments, workflows, comparisons, states and metaphors into clean diagrams that a reader can understand at a glance, and it is aimed at long-form articles, social image posts, blogs, Notion pages and spoken scripts.
- Is gimi-illustration-skill free to use? What license does it have?
- gimi-illustration-skill is open source under the MIT license. Check the license terms before using it in commercial products.
- Is gimi-illustration-skill still maintained?
- Updated within 6 months. The last commit to gimi-illustration-skill was on Aug 3, 2026.
- What can I use gimi-illustration-skill for?
- Adding illustrations to long-form Chinese articles or blog posts. Creating vertical image cards for Xiaohongshu-style social posts. Illustrating methodology, workflow or decision explanations for internal proposals. Producing storyboard-style visuals for spoken video scripts.
- How do I install gimi-illustration-skill?
- Clone the repository: git clone https://github.com/GiMi-Xiaomi/gimi-illustration-skill.git Enter the project folder: cd gimi-illustration-skill Copy the skill into the Codex skills folder: mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills" && cp -R . "${CODEX_HOME:-$HOME/.codex}/skills/gimi-illustration" In your chat, send 配图 followed by your article text
- Where can I get gimi-illustration-skill?
- The source code and installation instructions are on GitHub: https://github.com/GiMi-Xiaomi/gimi-illustration-skill
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An AI agent skill that turns an article into a 3:4 vertical cover image prompt for WeChat Official Accounts and Xiaohongshu, with optional face consistency.
90798OtherUpdated 3 months ago
A numbered library of 331 hand-drawn illustration styles, 167 layouts and 36 solid color themes that turns picks into bilingual prompts for AI image generators.
5.9K676HTMLOtherUpdated yesterday
An agent skill that makes AI coding agents report task results in controlled Simplified Chinese, based on ASD-STE100 principles.
32816MITUpdated 10 days ago
A collection of AI agent skills by Baoyu for content creation, image generation, and publishing to WeChat, X, and Weibo, for Claude Code, Codex, and similar agents.
26.8K2.9KTypeScriptMITUpdated last month
An AI agent skill set for Chinese patents: drafting invention disclosures and application documents, explaining patents, mapping patent landscapes, and assisting office action responses.
11.9K1.3KPythonMITUpdated 2 days ago