A Codex skill for clear, coherent scientific schematics and technical-route figures in grant applications, with particular attention to Chinese proposals reviewed by researchers outside the immediate specialty.
基金申请科研绘图 skill：核心概念一眼可认，科学逻辑清晰，画面紧凑，文字简洁，字体、框线和语义配色统一。
- Make the central research object or method immediately recognizable.
- Use terms and noun phrases instead of explanatory sentences inside the figure.
- Limit the figure to three font sizes; use bold labels, Chinese Heiti and Latin Arial.
- Use a white canvas, clear frames and selective fills for emphasis.
- Assign consistent colors to concepts across inputs, tools, details and outputs.
- Distinguish containment, tool calls, information flow, parallel branches, feedback and expanded views.
- Keep internal platform modules identifiable before expanding their details outside the main frame.
- Fit the canvas and label boxes to the content; do not add filler to match a reference figure’s density.
- Preserve scientific meaning and clearly distinguish schematic illustrations from measured results.
- Inspect the final editable source and its exported preview, including previously corrected issues, before delivery.
These are grant-figure design defaults, not universal publication rules. The user's instructions, scientific content and grant template take precedence.
Clone this repository into the local skill directory:
git clone https://github.com/Pyramtee/grant-figure-studio.git ~/.codex/skills/grant-figure-studioIf you use a custom skill directory, place the repository there instead. The installed folder contains SKILL.md at its root.
Use $grant-figure-studio to design a grant proposal figure from my approved research plan and reference figure. Discuss the layout first, then create an editable source and a PNG preview after I approve it.
使用 $grant-figure-studio，根据已确认的研究方案和参考图设计基金申请示意图。先讨论构图，确认后再绘制，并在交稿前检查逻辑、重点、密度、字体、线条和语义配色。
| File | Purpose |
|---|---|
| SKILL.md | Skill entry point and workflow |
| Design system | Typography, sizing, frames, palette and layout |
| Visual logic | Core concepts, platform containment, functional flows, expanded details, comparisons and feedback |
| Acceptance checklist | Final-source and export review |
| Textual examples | Agent and platform structures, parameters and transfer boundaries |
| UI metadata | Skill display name and description |
The skill instructions and detailed references are written in Chinese; the name and discovery description are in English. The skill provides design guidance, rather than a standalone rendering engine. Use the document or graphics tools available in your environment to produce the figure.
This public repository contains rules and textual examples only. Original reference artwork, PPTX files and project figure previews are kept outside the public package. Provide reference materials that you are authorized to use for each new task.