nano-banana-image-gen_skill

Overview

This skill generates images using Google Gemini's image model for high-quality, customizable visuals from text prompts. It supports optional reference images for style or character consistency and multiple aspect ratios. Use the included CLI or Python API to create single images or batch workflows for storyboards and comics.

How this skill works

The skill sends your text prompt (and any reference images) to Google Gemini's image model and downloads the resulting image to a specified output path. The CLI accepts prompt, aspect ratio, reference file(s), output path, and retry settings. A Python API supports batch generation and programmatic integration with pipelines.

When to use it

• Create single images from text prompts (illustrations, concept art, thumbnails).
• Maintain consistent style or character by providing reference images.
• Generate multiple related images (comic panels, storyboards) in a batch.
• Produce images with specific aspect ratios for social, print, or UI needs.
• Integrate image generation into automation or other agent workflows.

Best practices

• Provide a clear, descriptive prompt including style, mood, and composition for predictable results.
• Include reference images when you need consistent character or stylistic continuity.
• Choose an aspect ratio that matches your final use (1:1 for avatars, 16:9 for widescreen, 9:16 for vertical).
• Keep sensitive or policy-restricted content out of prompts to avoid failures.
• Check environment variables GOOGLE_API_KEY or GEMINI_API_KEY are set before running generation.

Example use cases

• Quick concept art: generate a set of thumbnail concepts for a product or scene.
• Character sheets: produce consistent poses or expressions using a reference image.
• Marketing assets: create social images in 1:1 or 16:9 aspect ratios.
• Comics/storyboards: batch-generate sequential panels with consistent style.
• Automation: call the Python API from other skills to embed image generation into pipelines.