banner-creator

The Banner Creator skill provides an automated, iterative pipeline for generating high-quality visual assets for developers and content creators. Designed to integrate directly into agentic workflows, it leverages the Gemini 3 Pro Image model (via the nanobanana skill) to produce custom imagery for various digital platforms. Users can start from conceptual requirements—specifying the purpose, branding, and style—and move through a structured process of batch generation, refined selection, and precise cropping to meet industry-standard aspect ratios like 2:1 for GitHub or 3:1 for Twitter headers.

• Automated batch generation of 20 unique image variations based on specific prompt engineering patterns.

• Iterative feedback loop allowing users to refine, re-generate, and polish favorite design directions.

• Integrated aspect ratio management and cropping tools for consistent branding across GitHub README, Twitter, LinkedIn, and website hero sections.

• Localized file management within the .skill-archive directory to track project history and ensure reproducibility.

• Interactive preview HTML generation for streamlined browser-based visual comparison and selection.

• Requires an active GEMINI_API_KEY and the nanobanana skill installed in the environment.

• Designed for use in collaborative agent environments like Claude Code, Cursor, or Windsurf.

• Best results are achieved by providing detailed style descriptions, specific color preferences, or existing logo assets for composition.

• The workflow prioritizes generation at a wide 21:9 ratio to ensure maximum flexibility for final crops.

• Output assets are organized by date and project summary to maintain a clean project structure throughout the design lifecycle.