figure-bake

What problem does it solve?

This Skill eliminates the tedious, error-prone manual work of designing, coding, reviewing, and formatting research figures for academic manuscripts, saving researchers hours of iterative tweaking and tooling coordination.

Core Features & Use Cases

• 3-agent iterative refinement loop: Automatically coordinates a figure designer, code implementer, and peer reviewer to iterate on a single figure until it meets publication standards, with no manual coordination required between steps.
• Dual artifact production: Renders a high-quality vector PDF for manuscript inclusion and a cropped outline-ready PNG for presentations, with automatic cropping to match venue formatting requirements.
• Resumable state tracking: Tracks all figure state, design briefs, render scripts, and critique history in the figures.json state file, so runs can be safely paused and resumed without losing progress.
• Use case: If you are writing a biology paper and need a multi-panel figure showing gene expression across treatment conditions, invoke this Skill with the figure ID and it will handle the full workflow from design brief to final cropped artifacts, ready to embed in your manuscript.

Quick Start

Invoke the figure-bake skill with your target figure ID (for example, /figure-bake fig1) to start the automated end-to-end figure creation workflow for that figure.