nano-banana-prompting-skill

What This Skill Does

The nano-banana-prompting-skill acts as an intelligent intermediary between user intent and Gemini 3 Pro Image generation. Most image models struggle with vague natural language requests, often defaulting to generic "AI-style" visuals. This skill elevates image generation by analyzing user requests to detect the intended artistic or photographic style, then programmatically assembling a structured JSON payload that includes specific technical parameters. By injecting industry-standard lighting, lens, and stylistic keywords into the prompt, the skill ensures that results are consistent with the desired look—whether that is a crisp product mockup, a cinematic photograph, or a stylized 3D render.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/minilozio/nano-banana-prompting Ensure you have the necessary dependencies for the nano-banana-pro directory path as defined in your system configuration.

Use Cases

This skill is designed for creators, developers, and marketers who require consistent visual assets. Use it for generating professional e-commerce product shots without a photography studio, creating concept art for game design, or producing high-quality marketing assets. It is particularly effective for teams that need to maintain a specific "brand look" across multiple images by leveraging its structured prompting capabilities.

Example Prompts

1. "Generate a realistic portrait of an elderly watchmaker working in his shop, cinematic lighting, 35mm lens, depth of field."
2. "Create a sleek product shot of a minimalist black smart speaker, studio lighting, soft shadows, white background, high-end e-commerce style."
3. "Draw an anime-style illustration of a cybernetic cat sitting on a futuristic Tokyo rooftop, Ghibli-inspired aesthetic, vibrant colors."

Tips & Limitations

• Contextual Precision: The more descriptive you are about the subject and environment, the better the skill can map your request to the appropriate technical keywords.
• File Naming: Always use descriptive names like product-shot-v1.png instead of generic names to keep your workspace organized.
• Reference Images: When providing reference images, ensure they are high-resolution for the best results.
• Limitations: The skill is optimized for Gemini 3 Pro; results may vary if used with other underlying models. It does not perform actual image editing, but rather guides the generation of new images based on provided references.