guizang-material-illustration
Guizang's material-style illustration Skill, turning concepts, processes and charts from your writing into explainer images with Chinese labels.
- Ecosystem
- Agent Skills
- Skill category
- Design
- Creator / organization
- 歸藏(GitHub:op7418)
What it is
guizang-material-illustration is an illustration Skill by Guizang (GitHub: op7418), who says it works in agent environments such as Claude Code and Codex. It focuses on the image in the middle of an article, slide or social card, drawing concepts, processes, mechanisms and system relationships as material-style explainer images with short Chinese labels, arrows and legends. Give it a messy chart screenshot or a set of numbers and it pulls out the chart type, axes, units and values, then redraws the chart so it reads faster. For obscure terms, scientific devices or historical objects, it looks up reference information first and then brings everything into the same style. The default theme color is IKB blue, with options such as lemon yellow and safety orange. The labels are part of the picture, which sets it apart from decorative illustrations. It belongs to the same series as dsh-guizang-design, the same author's plugin already listed on this site: the plugin's image cards handle the card's title and layout, and this Skill draws the illustration in the middle.
Who it's for:writers, presenters and teachers who need abstract ideas drawn as images people get at a glance.
Before you start: Install it with one npx skills add command, or have an agent with terminal access clone the repo into ~/.claude/skills/. Your client needs image generation; the author mentions GPT-Image, imagegen or the host's own image tools. When it looks things up, your material and reference images go to the image service you use, under that service's pricing and data terms. The docs are mostly in Chinese, and it can also make images in English. The README says it isn't meant for full card layouts, photo retouching, portraits, or research figures that need publication-grade precision. The repo doesn't state a license, so ask the author before reusing it.
Based on the project's public documentation. Not tested by AI Inspiration World.