AIA Academic Illustrator skill
A Codex skill turns a paper or method into a visual blueprint, then draws an architecture figure with built-in image generation.
- 468 ★ on the repository
- Last commit
Minimal access
Why this level
- The skill creates images and text files in the working folder.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/aia-academic-illustratorDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill creates images and text files in the working folder.
This is third-party code. Review the repository files before installing.
What it does
Next to the web app, the repository holds the Codex skill `academic-illustrator`. It takes a paper, PDF or method description, builds a visual blueprint with sections, labels and links, then creates the figure with Codex built-in image generation. No external model key is needed. The output is an image, `schema.md`, `render-prompt.txt` and a `qa.md` check. On request the agent hands the result to the `image-to-editable-ppt` skill and the figure becomes an editable `.pptx`. You can stop at the blueprint and review it before drawing.
Who it is for. Paper authors and PhD students who need a CVPR or NeurIPS level method figure without manual layout in an editor.
Good fit when
- A method figure is needed from the Method section, with labels matching the paper text.
- You want to review the visual blueprint first and only then spend generation on the picture.
- A finished picture needs a targeted fix, such as a label or an arrow.
- The figure must be delivered as an editable PowerPoint for further work.
Not a fit when
- The agent is not Codex and has no built-in image generation.
- A chart with exact numbers from data is needed: a plotting library fits better.
- The journal bans neural-network-generated images.
Example request
Use $academic-illustrator to draw an architecture figure from the Method section, show the visual blueprint first and wait for confirmation.Limitations
The skill targets Codex: copy `skills/academic-illustrator` into `${CODEX_HOME:-$HOME/.codex}/skills/`, and compare an existing skill of the same name first instead of overwriting it. Generation uses only the built-in tool and the skill does not fall back to external APIs on its own. PowerPoint output needs a separately installed `image-to-editable-ppt`. Check the text inside the generated picture against the paper, since image models mix up formulas. The documentation is written in Chinese.
How to disable. Delete the `skills/academic-illustrator` directory from the Codex skills folder. The web app in the repository is separate and unaffected by the skill.
FAQ
Is an external model key required?
No. Codex runs the analysis and its built-in generation creates the image. The web version keys are not needed by the skill.
What do I get?
An image plus `schema.md`, `render-prompt.txt` and `qa.md`. If PowerPoint is needed, a verified `.pptx` is added.
Can I see the plan first?
Yes, ask for the visual blueprint first and wait for your confirmation. The picture appears after your reply.
Does it work with an existing picture?
Yes. Modes include drawing from a reference and editing an existing figure, plus targeted fixes for flaws found during the check.
Is the skill tied to the repository web version?
No. The Next.js app and backend of the web version stay in the repository, but the skill offers no HTTP interface to the built-in model, and the two parts do not replace each other.
Discussion
No questions or reviews yet.
Related
SkillsA skill that turns an article, report or notes into a deck of full-slide images and assembles them into a pptx
A set of Codex skills for research work: image to editable Office shapes and VBA, and paper to a PPTX deck
Three skills for paper figures: fig-analyze reads material, fig-draw picks a style and renders, fig-pptx builds an editable PPT.
A Codex skill that rebuilds a slide from an image or PDF into PowerPoint where text, tables and arrows stay editable objects.
Four local skills for Codex: draw.io diagrams, interior plans, presentations and image work on a whiteboard.
Skill turns the agent into a deck producer: brief, storyboard, slide posters, QA and a PDF or editable PPTX.
README in short
The skill moves the earlier paper, blueprint, picture workflow into Codex and adds a check of the finished image and targeted fixes.