What problem does it solve?
Creating high-quality, conference-compliant academic figures (such as model architecture diagrams, method flowcharts, and conceptual illustrations) for research papers typically requires specialized graphic design skills and strict adherence to venue-specific visual standards. This Skill eliminates that manual effort by automating the generation of print-ready, grayscale-compatible figures that meet the requirements of top-tier conferences like CVPR, NeurIPS, ICLR, and MICCAI.
Core Features & Use Cases
- Native Codex Image Generation: Uses a local Codex app-server MCP bridge for high-quality raster rendering, with no manual bitmap fallbacks allowed.
- Strict Style Enforcement: Automatically validates figures against top-tier conference style guidelines, including clean typography, thick labeled arrows, restrained color palettes, and print-friendly formatting.
- Iterative Refinement Workflow: Runs a multi-stage review process with a minimum 9/10 quality score requirement, ensuring figures are logically clear, properly labeled, and paper-ready before finalization.
- Use Case: A researcher preparing a MICCAI 2025 paper on vertebrae segmentation can use this Skill to generate a publication-quality architecture diagram of their FMC-Net model, complete with labeled data flow arrows and LaTeX integration for seamless inclusion in their manuscript.
Quick Start
Use the paper-illustration-image2 skill to generate a publication-quality architecture diagram of the FMC-Net vertebrae segmentation pipeline with CVPR-style formatting, labeled data flow arrows, and a minimum 9/10 quality score.