This skill should be used when the user asks to "generate an image", "create an image", "optimize a prompt", "improve my prompt", "make my prompt better", mentions "nanobanana", "image prompt", "prompt engineering for images", or needs guidance on crafting professional AI image prompts. Provides the 6 core rules from analyzing 1,186 viral prompts.
From gemini-image-gennpx claudepluginhub aryanxpatel/aryanxpatel-plugins --plugin gemini-image-genThis skill uses the workspace's default tool permissions.
references/scene-guide.mdreferences/terminology.mdreferences/top-prompts.mdGuides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Generates FastAPI project templates with async routes, dependency injection, Pydantic schemas, repository patterns, middleware, and config for PostgreSQL/MongoDB backends.
Transform casual image descriptions into professional, high-quality prompts using patterns extracted from 1,186 viral AI-generated images.
Replace vague aesthetic words with specific professional terminology, proper nouns, brand names, or artist names.
| Instead of... | Use... |
|---|---|
| Cinematic, vintage | Wong Kar-wai aesthetics, Saul Leiter style |
| Film look, retro | Kodak Vision3 500T, Cinestill 800T |
| Warm tones | Sakura Pink, Golden Hour warmth |
| Japanese style | Wabi-sabi aesthetics, MUJI visual language |
| High-end design | Swiss International Style, Bauhaus functionalism |
Key terminology banks:
Replace subjective adjectives with specific technical parameters.
| Instead of... | Use... |
|---|---|
| Professional looking | 90mm lens, f/1.8, high dynamic range |
| From above | 45-degree overhead angle |
| Soft lighting | Soft side backlight, diffused light |
| Blurred background | Shallow depth of field, f/1.4 bokeh |
| Dramatic | Volumetric light, chiaroscuro lighting |
| Wide shot | 16mm wide-angle lens |
Explicitly state what NOT to include to prevent unwanted elements.
Common constraints:
Go beyond visual descriptions by adding multiple sensory dimensions.
Sensory layers:
For complex scenes, organize information into logical groups.
Standard grouping pattern:
[Subject Description]
Visual Style:
[Aesthetic references, color palette, artistic style]
Lighting & Atmosphere:
[Light sources, mood, environmental conditions]
Technical Parameters:
[Lens, aperture, film stock, resolution]
Constraints:
[What to avoid, what to preserve]
Choose format based on complexity:
Consult references/scene-guide.md for detailed patterns by scene type:
When generating an image, follow this intelligent flow:
Analyze the user's request and classify into one of these genres:
| Genre | Trigger Keywords |
|---|---|
food | dish, meal, cuisine, recipe, ingredient, cooking |
portrait | person, face, woman, man, selfie, headshot |
product | product, packaging, brand, commercial, advertising |
3d | icon, render, 3D, isometric, emoji, character |
cinematic | movie, scene, dramatic, action, poster |
design | UI, app, poster, layout, typography, mockup |
Read ${CLAUDE_PLUGIN_ROOT}/data/prompts-by-category.json and extract techniques from top prompts in that genre.
Top techniques by genre:
User input: "a bowl of ramen"
Optimized output:
Steaming bowl of authentic Japanese tonkotsu ramen, rich milky pork bone broth.
Visual Style:
High-end culinary magazine aesthetic. Warm earth tones with cream and amber highlights. Shot with Hasselblad medium format quality.
Composition:
45-degree overhead angle, 85mm lens, f/2.8, shallow depth of field on the soft-boiled egg.
Sensory Details:
Steam wisps rising and curling, noodles glisten with broth, chashu pork with caramelized edges, the aroma of garlic and sesame seems to penetrate the frame.
Constraints:
No utensils in frame. No text or watermarks. Maintain appetizing warm color temperature.
For detailed patterns and terminology, consult:
references/scene-guide.md - Scene-specific optimization patternsreferences/terminology.md - Complete professional terminology banksreferences/top-prompts.md - Top 50 viral prompts for inspirationExecute image generation:
node "${CLAUDE_PLUGIN_ROOT}/scripts/gen.js" "YOUR_OPTIMIZED_PROMPT"
This skill is optimized for Gemini 3 Pro Image (gemini-3-pro-image) via the antigravity-claude-proxy at localhost:8080.
The model excels at: