nanaban

The nanaban skill acts as a bridge between your AI agent and the Gemini image generation models, enabling the creation and manipulation of visual assets directly from the terminal. Designed for users who need a seamless workflow for graphic generation, this skill manages the complexities of the Gemini API through a straightforward CLI interface. Whether you are a designer building rapid prototypes, a developer creating social media assets, or a content creator needing specific visual styles, this tool automates the interaction cycle from prompt refinement to local file retrieval.

• Generates diverse visual media including illustrations, logos, icons, avatars, banners, and pixel art using Nano Banana and Imagen 4 models.

• Performs complex image transformations such as background removal, style transfer, color correction, restoration, and professional retouching.

• Supports specific aspect ratio adjustments (16:9, 9:16, 1:1) and high-resolution output (2K/4K) to meet precise design requirements.

• Facilitates iterative design through cost-efficient model selection, allowing users to balance quality and speed based on project budgets.

• Integrates directly with the host terminal environment for image previews and file management, ensuring generated assets are immediately available for review.

• Requires a valid Gemini API key and the installation of the nanaban CLI tool via the provided repository.

• Use flash models for general generation and editing tasks; reserve pro models for high-quality, professional-grade requirements.

• Use the generate command with detailed prompts covering subject, style, lighting, and composition for optimal results; iterate based on previous outputs.

• Leverage the edit command to modify existing local images; note that editing is limited to Nano Banana models while Imagen 4 is optimized for initial generation.

• Trigger this skill for creative visual tasks rather than technical image processing like file optimization or programmatic resizing.