What problem does it solve? Creating images through AI generation often produces poor results when prompts are written as keyword lists or lack structure. This Skill provides best-practice prompting guidance and a clear workflow for using the nano-banana-2 MCP image generation tools, so generated images match the intended composition, style, and text content. ## Core Features & Use Cases - Narrative Prompting Guidance: Enforces paragraph-style prompts that achieve roughly 94% coherence versus 61% for keyword lists, with rules for layout, text length, and style. - Generation and Iterative Editing: Covers the full workflow of generating an image, reviewing the saved file, and refining it with continue_editing calls. - Configuration and Defaults: Provides sensible defaults for resolution, aspect ratio, thinking mode, and output format, plus a first-time setup check for the GEMINI_API_KEY. - Use Case: A user asks for a side-by-side comparison diagram with labeled arrows. The Skill guides the model to write a narrative prompt, set thinking to high for the complex layout, generate the image, and refine labels through follow-up edits. ## Quick Start Ask the assistant to generate an image, for example: create a flat-style diagram showing the four stages of a construction project lifecycle with labeled arrows.
