Plan illustrations with AI
Choose an illustration’s direction, meaningful parts and constraints before production, then hand over a readable plan rather than a premature image.
Plan illustrations with AI before committing to a drawing. The illustration-plan skill turns a brief into visual decisions: where the work will appear, how small it will be seen, which parts matter and what must stay unchanged.
The deliverable is a choice of directions or a semantic scene plan. The plan describes the drawing. An illustrator or another tool can use it to create the artwork. A planning-only request ends with the plan; a production workflow can use that handover when you choose to continue.
What it does
- Compare distinct visual directions before production
- Prepare a concise handover for an illustrator
- Record the purpose of important scene elements
- Keep approved branding and wording unchanged
- Balance small-size readability against surface detail
- Separate inspiration from a faithful reconstruction brief
Install and use
- Install the complete folder in your agent’s skill directory, or select this leaf through the maker’s linked Skills CLI route. Use a file-capable agent; structural validation needs Python 3.10 or later, without a renderer or model API.
- Try: “Suggest three directions for a small rocket in an onboarding guide. It will appear at 96 pixels, must retain our blue palette, and should remain a plan for now.” Supply authorised references and fixed wording.
- Choose a direction. The agent records the canvas, meaningful IDs and roles, palette hints, constraints and consequential assumptions in scene.json, leaving geometry to production.
- Check the plan yourself and use its validator for the document contract. A successful check does not establish design quality or verify a future SVG.
FAQ
Source
Made by Yein Sung. Original skill: illustration-plan, published on GitHub under the MIT licence.
View on GitHub