SVG Scientific Figure Generation

Generate publication-quality, editable SVG scientific illustrations through a Review-Refine loop. Unlike the Gemini image model approach (PNG, non-editable), this skill produces vector SVG code that users can modify in draw.io, Adobe Illustrator, Inkscape, or PowerPoint.

When to Use

• User needs a mechanism diagram (signaling pathway, cellular process, drug mechanism)
• User needs an experimental workflow figure (study design, analysis pipeline)
• User needs a conceptual figure (graphical abstract, model summary)
• User explicitly asks for editable or SVG figures
• The existing Gemini diagram skill generated something that needs precise label/layout control

When NOT to use (use other approaches):

• Data-driven plots (boxplot, volcano, KM curve) → use Python/R code
• Microscopy/imaging results → use Gemini image generation
• Quick sketches → use Gemini image generation

Review-Refine Architecture