From AI Image Generation
Expert prompt engineering guide for photorealistic and creative image generation using Nano Banana Pro / Gemini Image Ultra. Includes camera specs, lighting, film aesthetics, and portrait prompts.
How this skill is triggered — by the user, by Claude, or both
Slash command
/image-gen:nano-banana-proThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
> **See Also:**
See Also:
- image-generation - General prompt engineering for all image generators
- gemini-3-pro - Full Gemini suite: Imagen 3, Veo 2, TTS, Live API
- openai-dalle - OpenAI suite: DALL-E 3, Sora 2, Whisper
This skill provides expert-level prompts for photorealistic and creative image generation using Nano Banana Pro (Gemini Pro Image Ultra).
Ultra-sharp, full-color large-format image shot with [Camera] and [Lens]
Best Cameras:
Best Lenses:
Three-point lighting setup:
- Key light at 45 degrees (warm, 5600K)
- Fill light at -30 degrees (soft, 50% intensity)
- Rim light from behind (golden, highlights hair/shoulders)
- Vignette effect for focus
Lighting Terms:
Shot on [Film Stock], [ISO], natural film grain
Film Stocks:
{
"subject": {
"age": 35,
"gender": "female",
"expression": "confident smile, eyes engaged",
"skin": "natural texture with visible pores, subtle makeup",
"hair": "styled professionally, individual strands visible"
},
"photography": {
"camera": "Sony A7III",
"lens": "85mm f/1.4",
"aperture": "f/2.0",
"lighting": "soft natural light from large window, reflector fill"
},
"style": {
"background": "clean gradient, subtle bokeh",
"color_grade": "neutral with slight warmth",
"quality": "8K resolution, ultra-sharp focus on eyes"
},
"preserve_original": true
}
Golden Hour portrait of [subject], shot on Kodak Portra 400
Natural backlight creating golden rim around hair
Soft catchlights in eyes, natural skin texture
Expression: [emotion] - subtle, genuine
Background: [setting] with beautiful bokeh
Film grain: subtle, organic
{
"era": "2000s",
"setting": "bathroom mirror selfie",
"subject": {
"clothing": ["low-rise jeans", "baby tee", "chunky belt"],
"accessories": ["flip phone", "dangly earrings", "butterfly clips"],
"pose": "classic mirror selfie angle"
},
"aesthetics": {
"quality": "early digital camera, slight blur",
"flash": "direct flash, harsh shadows",
"color": "slightly oversaturated",
"timestamp": "bottom right corner"
}
}
Aerial view of massive crowd gathered in [location]
Thousands of people, each with distinct features and clothing
"Where's Waldo" style complexity
Every person has unique appearance, no repetition
Photorealistic quality, natural lighting
Photo series of same person aging through the years:
- Age 20: [description]
- Age 40: [description]
- Age 60: [description]
- Age 80: [description]
Keep facial features exactly consistent across all ages
Same bone structure, eye shape, unique characteristics
Image containing itself recursively (Droste effect)
[Subject] holding a frame showing the same scene
Infinite regression, each iteration smaller but detailed
Mathematical precision in the recursion
Photograph taken at exact coordinates:
Latitude: [X], Longitude: [Y]
Time: [HH:MM], Date: [YYYY-MM-DD]
Season: [season]
Weather conditions: [weather]
Capture the atmosphere and location authentically
"How [profession] sees [object]"
Example: "How engineers see the Golden Gate Bridge"
- Show technical annotations
- Structural analysis overlays
- Material specifications
- Force vectors and stress points
ALWAYS include in prompts for consistent faces:
- "Keep facial features exactly consistent"
- "preserve_original: true"
- "Same bone structure, eye shape, nose, lips"
- Reference specific features to maintain
Request micro-details:
- "Individual hair strands visible"
- "Natural skin pores and texture"
- "Fabric fibers and weave pattern"
- "Reflections in eyes showing light source"
- "Subtle imperfections for realism"
Resolution & Quality keywords:
- "8K resolution"
- "Ultra-sharp"
- "RAW quality"
- "Uncompressed"
- "Professional retouching"
- "Magazine quality"
For maximum control, use JSON structure:
{
"image_type": "portrait | landscape | product | abstract",
"subject": {
"description": "detailed subject description",
"age": "approximate age if person",
"expression": "emotional state",
"clothing": ["item1", "item2"],
"accessories": ["item1", "item2"],
"pose": "body position description"
},
"environment": {
"setting": "location description",
"time_of_day": "golden hour | midday | night | etc",
"weather": "sunny | overcast | rain | etc",
"background": "background details"
},
"photography": {
"camera": "camera model",
"lens": "lens specification",
"aperture": "f-stop value",
"shutter_speed": "if relevant",
"iso": "ISO value",
"film_stock": "if film aesthetic"
},
"lighting": {
"type": "natural | studio | mixed",
"key_light": "main light description",
"fill_light": "fill light if any",
"rim_light": "rim/back light if any",
"color_temperature": "warm | neutral | cool"
},
"style": {
"color_grade": "color palette description",
"mood": "emotional tone",
"quality": "resolution and sharpness",
"post_processing": "editing style"
},
"constraints": {
"preserve_original": true,
"avoid": ["elements to avoid"],
"emphasis": ["elements to emphasize"]
}
}
Professional product photography of [product]
White seamless background, soft diffused lighting
Multiple catch lights for dimension
Sharp focus throughout, no shadows on background
8K resolution, color-accurate
Instagram-worthy portrait of [subject]
[Setting] with aesthetic background
Natural posing, candid feel
Soft editing, skin smoothing (subtle)
Square format, centered composition
Warm color grade, slight lift in shadows
High-fashion editorial shot for [magazine]
Model wearing [outfit] in [setting]
Dramatic lighting, strong shadows
Bold color story: [colors]
Shot by [photographer style reference]
Full-length, dynamic pose
Architectural photography of [building/interior]
Wide-angle lens (24mm), straight verticals
HDR technique, balanced exposure
Blue hour exterior / natural light interior
Clean, aspirational aesthetic
Remove distracting elements
Use this template to restore old/blurry/compressed photos while preserving identity, pose, and composition. Works best with Nano Banana Pro / Gemini 3 Pro Image.
ДИРЕКТИВА: Выполнить реставрацию ультра-верности и кинематографический
перерендеринг изображения-источника. Главная цель — «Аналитическая
Реконструкция Микро-Деталей» на основе источника низкого качества.
СУБЪЕКТ: [описание основного субъекта — человек/объект/животное]
КРИТИЧЕСКИ ВАЖНО: Структура лица, идентичное выражение и точная поза
должны СТРОГО СОВПАДАТЬ с предоставленным изображением-источником.
Требуется 100%-ная верность идентичности.
РЕКОНСТРУИРУЕМЫЕ ДЕТАЛИ:
- Экстремальное восстановление текстур кожи (поры, микро-морщины)
- Реалистичные индивидуализированные пряди волос
- Кристально чистые глаза с чёткими бликами
- Чистые и чётко очерченные края одежды и аксессуаров
СЦЕНА: Точная копия оригинальной композиции.
КАДРИРОВАНИЕ: [Close-up | Medium Shot | Full Body], Правило Трети.
ОКРУЖЕНИЕ: Реконструкция оригинального фона. Увеличить глубину резкости
для разделения субъекта и окружения, устранить артефакты и шум.
ТЕХНИЧЕСКИЕ ХАРАКТЕРИСТИКИ:
- КАМЕРА: Объектив 85 мм, апертура f/1.8 (bokeh)
- ОСВЕЩЕНИЕ: Сбалансированное Кинематографическое (Rembrandt Lighting),
высокий контраст, детали в тенях сохранены, точные зеркальные блики
- КАЧЕСТВО: 8K, Фотореалистичные Текстуры, ProRes, Студийная Чёткость
ОРИЕНТИР: Прикрепленное изображение — ЕДИНСТВЕННЫЙ источник структуры
и идентичности. НЕ додумывать черты.
НЕГАТИВНЫЙ ПРОМПТ: Размытость, нечёткость, визуальный шум, артефакты
сжатия, изменение идентичности, изменение выражения, изменение позы,
изменение одежды, добавление объектов, перерисовка, вид рисунка или
картины, низкое разрешение, AI-look, uncanny valley.
DIRECTIVE: Execute ultra-fidelity restoration and cinematic re-render of
the source image. Primary goal — "Analytical Micro-Detail Reconstruction"
from the low-quality source.
SUBJECT: [main subject description]
CRITICAL: Facial structure, identical expression, and exact pose must
STRICTLY MATCH the source image. 100% identity fidelity required.
RECONSTRUCTED DETAILS:
- Extreme recovery of natural skin textures (visible pores, micro-wrinkles)
- Realistic individualized hair strands
- Crystal-clear eyes with defined catchlights
- Clean, well-defined edges on clothing and accessories
SCENE: Exact copy of original composition.
FRAMING: [Close-up | Medium Shot | Full Body], Rule of Thirds.
ENVIRONMENT: Reconstruct original background. Increase DOF for subtle
but clear subject/environment separation. Eliminate artifacts and noise.
TECHNICAL SPECS:
- CAMERA: 85mm lens, f/1.8 aperture (soft bokeh)
- LIGHTING: Balanced Cinematic Lighting (Rembrandt), high contrast,
preserved shadow detail, precise specular highlights
- OUTPUT QUALITY: 8K resolution, photoreal textures, ProRes quality,
studio clarity, poster-grade realism
REFERENCE: Attached image is the ONLY reference for structure and
identity. Do not invent features.
NEGATIVE PROMPT: Blur, softness, visual noise, compression artifacts,
identity change, expression change, pose change, clothing change,
added objects, repainting, painterly look, low resolution, AI-look,
uncanny valley.
| Input | Best For |
|---|---|
| Old family photos | Restoration with identity preservation |
| Low-res screenshots | Upscaling with detail recovery |
| Compressed JPEGs | Artifact removal + micro-detail recovery |
| Scanned prints | Texture reconstruction, scratch removal |
| Blurry phone photos | Sharpness + studio-grade re-render |
Portrait close-up: Change КАДРИРОВАНИЕ to Close-up (поясной план), camera to 135mm f/2.0.
Full-body restoration: Full Body, camera 35mm f/2.8, lighting Natural Soft Light.
B&W → Color: Add Колоризация: исторически точные цвета эпохи [1950s/1980s/etc.].
Group photo: Add Сохранить идентичность КАЖДОГО человека отдельно, без смешения черт.
| Goal | Key Terms |
|---|---|
| Sharp portrait | 85mm f/1.4, eye focus, catchlights |
| Cinematic | 35mm, shallow DOF, film grain, color grade |
| Natural | Golden hour, soft light, candid pose |
| Professional | Studio lighting, clean background, sharp |
| Vintage | Film stock, grain, era-specific styling |
| Dramatic | High contrast, Rembrandt lighting, shadows |
| Soft/Dreamy | Backlit, overexposed background, pastel |
npx claudepluginhub jhamidun/claude-code-config-pack --plugin image-genBuilds a throwaway prototype to answer a design question about UI appearance or state/logic behavior. Guides you through two branches: interactive terminal app for logic validation, or multiple UI variations for visual exploration.