From model-bridge
Generate optimized prompts for Google Imagen 4 models. Teaches concrete visual language, composition techniques, and model-specific patterns. Use when creating images with Imagen, Google AI image generation, or Vertex AI imaging.
How this skill is triggered — by the user, by Claude, or both
Slash command
/model-bridge:imagen-promptThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Craft effective prompts for Google Imagen 4 (Ultra/Fast) that produce high-quality images. Teaches concrete visual language, composition patterns, and the difference between abstract concepts and specific imagery that Imagen understands.
Craft effective prompts for Google Imagen 4 (Ultra/Fast) that produce high-quality images. Teaches concrete visual language, composition patterns, and the difference between abstract concepts and specific imagery that Imagen understands.
vertex-setup - For configuring Google Cloud and Vertex AI accessflux-prompt - Alternative for local image generationUse this skill when the user wants to:
Imagen 4 Ultra (imagen-4.0-ultra-generate-001):
Imagen 4 Fast (imagen-4.0-fast-generate-001):
Critical Insight: Imagen 4 does not support inpainting or iterative editing. Invest in prompt quality upfront - single-shot generation with Ultra produces the best results.
For straightforward images with one main focus:
[SUBJECT with specific details], [SETTING/BACKGROUND], [LIGHTING], [STYLE]
Example - Product Shot:
Sleek wireless headphones in matte black finish, floating against a gradient
background transitioning from deep blue to purple, soft studio lighting with
subtle reflections, product photography style, clean and minimal
Example - Portrait:
Young woman with curly red hair and green eyes, wearing a cream knit sweater,
sitting by a rain-streaked window, soft natural light from the left,
candid portrait photography, warm color tones
Example - Food:
Fresh pasta dish with basil and cherry tomatoes in a white ceramic bowl,
rustic wooden table surface, steam rising, warm overhead lighting,
food photography style, shallow depth of field
For images with multiple elements that need specific placement:
[POSITION]: [Element with details]
[POSITION]: [Element with details]
[POSITION]: [Element with details]
[Overall style and lighting]
Positions: LEFT, RIGHT, CENTER, FOREGROUND, BACKGROUND, TOP, BOTTOM
Example - Scene with Multiple Elements:
LEFT: Vintage red bicycle leaning against a brick wall, basket with
fresh flowers. CENTER: Cobblestone alley leading to a sunlit courtyard.
RIGHT: Small cafe with outdoor seating, striped awning.
Golden afternoon light, European street photography style
Example - Product in Context:
FOREGROUND: Leather messenger bag in cognac brown, buckles catching light.
CENTER: Person walking through busy city street, motion blur.
BACKGROUND: Glass storefronts reflecting afternoon sun.
Street photography style, shallow depth of field on bag
Imagen handles text well with specific formatting:
[Scene description] with the text "[YOUR TEXT]" in [font style/placement]
Example:
Vintage travel poster showing the Eiffel Tower at sunset, with the text
"PARIS" in bold Art Deco typography at the bottom, warm orange and gold
color palette, retro illustration style
Premium watch with rose gold case and black leather strap, positioned at
45-degree angle on dark slate surface, dramatic side lighting creating
sharp reflections on metal, product photography, luxurious feel
Misty mountain lake at dawn, snow-capped peaks reflected in still water,
pine forest silhouettes on the shore, soft pink and orange sky,
landscape photography, wide 16:9 composition
Massive ancient tree with glowing amber crystals embedded in bark,
bioluminescent mushrooms at base, mystical forest clearing with
shafts of moonlight, fantasy illustration style, rich jewel tones
Modern minimalist house with floor-to-ceiling windows, cantilevered
over rocky coastline, dramatic sunset behind, waves crashing below,
architectural photography, clean lines and geometric forms
Chef in white coat standing in professional kitchen, arms crossed,
stainless steel equipment behind, warm overhead lighting,
environmental portrait photography, confident expression
Instead of → Write:
| Abstract | Concrete |
|---|---|
| "beautiful sunset" | "orange and pink sky with scattered clouds" |
| "cozy room" | "warm lamp light, knit blanket on leather armchair" |
| "delicious food" | "steam rising, golden crust, fresh herbs visible" |
| "professional" | "crisp white shirt, clean workspace, natural light" |
| "scary atmosphere" | "deep shadows, single flickering light, fog" |
| "luxury" | "polished marble, gold accents, soft velvet" |
When user requests an Imagen prompt:
Clarify the subject - What's the main focus?
Determine complexity:
Add concrete details:
Choose ONE style anchor:
Specify aspect ratio based on use:
API configuration (for reference):
model='imagen-4.0-ultra-generate-001'
number_of_images=4 # Generate variations
aspect_ratio='16:9' # Or '1:1', '9:16', '4:3'
Elements not appearing: Be more specific about placement. Use spatial positions explicitly.
Wrong style: Use only ONE style anchor. Multiple styles confuse the model.
Chaotic output: Too many elements. Simplify to 3-5 key elements maximum.
Vague results: Replace abstract words with concrete descriptions.
Text not rendering: Put text in quotes and specify font style/placement clearly.
Start Simple: Begin with subject + lighting + style. Add complexity only if needed.
Be Specific: "Labrador retriever" beats "dog". "Mahogany desk" beats "wooden desk".
Light Matters: Describe direction, quality, and color of light.
One Style: Pick one clear style anchor and commit to it.
Generate Multiple: Always request 4 variations - quality varies between outputs.
Aspect Ratio: Match ratio to your use case (social, web, print).
npx claudepluginhub lando-labs/claude-plugins --plugin model-bridgeEnhances image generation prompts with Subject-Context-Style structure, lighting physics, camera terminology, and character consistency patterns. Useful for creating detailed, physically coherent image prompts.
Generates images from text, edits images with references, performs product placement, style transfer, and multi-image composition using OpenAI DALL-E or Google Gemini.
Generates optimized prompts for Gemini 2.5 Flash Image (Nano Banana) using best practices for photorealistic shots, art styles, and multi-turn editing workflows.