stem-illustration skill
An image skill for science figures: 24 templates for diagrams and posters in biology, chemistry, physics and maths.
- 137 ★ on the repository
- Last commit
Extended access
Why this level
- Prompts and the API key go to an external image-generation service.
- The skill runs scripts and saves images to disk.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/stem-illustrationDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Prompts and the API key go to an external image-generation service.
- The skill runs scripts and saves images to disk.
This is third-party code. Review the repository files before installing.
What it does
The skill assembles a request for a science illustration: the agent clarifies the discipline, diagram type, audience, terminology, size and language, picks one of 24 templates and returns a visual brief with a ready prompt. After confirmation it runs `scripts/generate_image.py`. Templates cover signalling pathways, experiment workflows, cells, posters, timelines, plasmid maps, mechanisms, physics schematics, chemical structures and model architectures. Four styles are available: academic, textbook, infographic and 3d-render. For strict geometry `scripts/generate_geometry.py` draws figures with matplotlib.
Who it is for. Researchers, teachers and technical writers who need a conceptual diagram or poster without drawing it by hand.
Good fit when
- A paper needs a journal-style signalling pathway scheme in a vertical layout.
- A lecture lacks an illustration of a principle such as buoyancy or cell division.
- A set of related figures must share one style: palette, fonts, line weights.
- A talk needs a bilingual poster in two languages.
Not a fit when
- You need a chart of real experimental data; the skill declines, use GraphPad, Origin or matplotlib.
- You want photorealistic shots or characters; use a general image-generation skill.
- The figure must show genuine gels, blots or micrographs; the skill does not fabricate them.
Example request
Draw a MAPK-ERK signalling pathway in a journal style, vertical, with English labels. Show the brief first.Limitations
Generation needs your own image-model access: set `IMG_BASE_URL`, `IMG_MODEL` and `IMG_API_KEY` in `.env`; the script supports an OpenAI-compatible synchronous API and the asynchronous apimart.ai API. A model can draw a wrong reaction or geometry: the author marks the aspirin synthesis sample in the README as erroneous. A subject expert should check the result. The repository states no license. Documentation and templates are mostly Chinese.
How to disable. Delete the skill directory from the agent's skills directory. Remove the `.env` file with the key and the `generated-images` folder separately.
FAQ
Which sizes does the skill recommend?
4:3 at 2K for a single journal column, 16:9 for two columns, 3:4 for vertical schemes, 2:3 for posters, 16:9 for slides and 21:9 for long flows.
How do I run generation manually?
Run `python scripts/generate_image.py --prompt-file ./prompt.txt --size 3:4 --resolution 2k`. For the synchronous API use `--prompt` with `--size 1024x1536 --quality high`.
Why does the skill refuse to draw data charts?
An academic integrity rule: it does not create statistical charts or experiment scans. Use GraphPad, Origin or matplotlib for data.
How do I keep one style across a series?
Multi-panel work applies Campaign Style Lock: shared palette, fonts, line weights and label boxes.
Which styles are available?
Four: `academic`, `textbook`, `infographic` and `3d-render`. The style is chosen together with the template when the brief is composed.
Discussion
No questions or reviews yet.
Related
SkillsA skill for scientific papers: vector schematics as SVG and PDF plus the code that draws them, so any element stays editable.
The skill turns paper text, methods and captions into structured prompts for scientific figures, tuned to the chosen image model.
A skill hand-writes SVG feature illustrations by one SPEC.md ruleset, with a single palette serving both light and dark themes.
A skill that draws interactive isometric illustrations in SVG, with WebGL shaders and React output, explaining an idea as a working machine
A skill that assembles ready image prompts in hand-drawn styles from a set of tested recipes: kids drawing, minimal line, crayon, Ghibli look and more
A skill that turns your photo or description into a hand-drawn character and uses it for diagrams, knowledge cards and slides
README in short
The repository holds the skill rules, 24 JSON templates, finished image samples and ten prompt rules: colours as `#RRGGBB` codes, explicit directions, standard terms and a negative list. The ban on fabricated data is described separately.
README badge
Original SKILL.md
--- name: stem-illustration description: STEM 领域 AI 图像生成 Skill。面向科研、教育、工程场景，生成信号通路、实验流程、机制图、细胞结构、概念信息图、学术海报、架构图等示意图。 --- # stem-illustration 面向 STEM（科学、技术、工程、数学）领域的 AI 图像生成 Skill，专注生成**科研示意图、教学插图、技术架构图**等概念性图像，覆盖生物医学、化学、物理、工程、数学 6 大学科。 ## 何时使用 - 用户要生成科研示意图、机制图、流程图、流程海报、教学插图、学术图、技术架构图 - 用户提供主题/论文/概念/数据，期望出一张图（不限于"科研"，包括科普、教学、产品技术图） - 学科关键词触发：信号通路、pathway、mechanism、细胞结构、anatomy、实验流程、workflow、质粒、plasmid、机器学习、neural network、力学、电路、化学结构、数学概念、信息图、海报、截面、爆炸图 ## 不适用 - 真实实验数据图（散点/柱状/线图，学术诚信）→ 拒绝并提示使用 GraphPad/Origin/Python matplotlib - 照片级写实（人像/风景/产品摄影）→ 用通用生图 skill - 创意插画/卡通角色 → 用通用生图 skill ## 生图配置 复用 [scripts/generate_image.py](scripts/generate_image.py)，自动兼容 OpenAI 同步 API 和 apimart.ai 异步 API。 `.env` 字段： | 字段 | 必填 | 说明 | |---|---|---| | `IMG_BASE_URL` | 是 | API 根地址 | | `IMG_MODEL` | 是 | 图片模型 | | `IMG_API_KEY` | 是 | API key | 模板：[.env.example](.env.example)（含 STEM 推荐参数说明）。 ## 推荐尺寸 | 用途 | 比例 | 分辨率 | |---|---|---| | 期刊单栏图 | 4:3 | 2K | | 期刊双栏图 | 16:9 | 2K | | 垂直机制图（信号通路/树） | 3:4 | 2K | | 学术海报（A0/A1） | 2:3 | 2K | | PPT/幻灯片插图 | 16:9 | 2K | | 正方形示意图 | 1:1 | 2K |