image-generation_skill

Overview

This skill generates professional-quality images using the FAL.ai nanobanana pro models. It guides prompt construction, model selection, and integration with automation so you can produce product shots, social graphics, brand assets, and other visuals directly in Claude Code. The focus is predictable, high-quality outputs and fast iteration.

How this skill works

You build prompts with a five-part formula (subject, description, style, technical details, mood) to control the model. The skill maps those prompts to the nanobanana pro family, tuning size, guidance scale, and inference steps for the right tradeoff between speed and fidelity. It integrates with an automation API so Claude Code can call generation endpoints, save assets to folders, and return file paths for downstream use.

When to use it

• Creating product photography or hero images for marketing and e-commerce
• Generating social media graphics and platform-specific posts quickly
• Producing logos, icons, or simple brand illustrations
• Batch-creating consistent assets for a campaign or multi-product launch
• Iterating visual concepts during design reviews or creative sprints

Best practices

• Start with a documented style guide (Creative Strategist) and reuse its descriptors
• Use the 5-part prompt formula to include subject, visuals, style, technical specs, and mood
• Generate multiple variations and iterate with adjusted guidance scale and steps
• Choose image size based on final use: 1024x1024 for product shots, 768x768 for social
• Organize outputs into named folders and include descriptive file names for assets

Example use cases

• Generate three 1024x1024 product photos of a watch with studio lighting and rim light
• Create a set of 1080x1080 Instagram images with a trending vibrant aesthetic
• Produce a minimalist tech logo in blue and white for immediate design review
• Batch-generate product photos for multiple SKUs to populate an e-commerce catalog
• Iterate hero images for a landing page using higher inference steps for quality