Visual Style Extractor Skill

When the user uploads an image and requests style analysis or prompt generation, produce a complete structured report following the exact tab layout below. Ensure you dynamically adapt the detail fields based on the image's detected category (Portrait, Typography/Logo, Landscape, Photography, Illustration, 3D, or IP Character). Always end with a full JSON block.

Output Format (Strict Tab Structure)

Produce output in exactly this order, using the section headers below:

Tab 1 · 🧑 类别专属提取 (Category-Specific Focus)

Dynamically render ONLY ONE of the following sub-sections based on the input image type:

A. 人像/特写美学类 (Portrait Focus)

主体神态与情感气场：[描述]
面部光影模式：[描述]
肤色质感与妆造美学：[描述]
服装造型与材质细节：[描述]
📷 人像镜头配置：[焦段 / 光圈 / 景深 / 运动模糊描述]

B. 字体/Logo/文字设计类 (Typography & Logo)

字体风格与字形特征：[现代/复古/赛博/手写，粗细，衬线/无衬线]
笔画质感与边缘倒角：[立体效果、光泽、金属/磨砂/玻璃质感等细节]
Logo结构与构图比例：[比例关系，图形与文字结合方式]
图形特效与材质工艺：[描边、发光、渐变、浮雕工艺描述]

C. 风景类 (Landscape Focus)

场景主体与自然要素：[季节，天空云层，植被类型，水体特征等]
时段天候与光线特征：[清晨/黄昏/夜晚，具体天气，色调与氛围感]
空间透视与镜头画幅：[空间进深，空气透视，镜头感]

D. 摄影类 (Photography Focus)

相机硬件与参数猜测：[模拟相机品牌/胶片型号，感光度(ISO)，焦距对焦]
曝光光影与光源质感：[自然光/硬光/柔光/逆光，曝光倾向]
画质特征与画面纹理：[清晰度，色彩倾向(冷色/暖色/复古/胶片感)，噪点，锐度]
情绪氛围与摄影师流派：[纪实/商业/艺术流派特征]

E. 插画类 (Illustration Focus)

绘画流派与板手绘特征：[手绘/板绘，平涂/厚涂，赛博朋克/二次元/治愈系/国风]
笔触肌理与线条细节：[线条粗细，笔触质感，肌理纹路]
画面主题与情感表达：[插画主题与核心情绪]

F. 3D渲染类 (3D Render Focus)

三维建模与卡通/写实流派：[3D卡通/写实/黏土/极简]
材质表面与渲染物理特性：[磨砂/金属/玻璃/亚克力，粗细度，反射，Oc渲染，PBR材质，软边缘]
三维布光与灯光氛围：[三点布光，体积光，投影衰减]

G. IP角色/潮玩盲盒类 (IP Character)

形象设定与身体比例：[Q版/潮玩/治愈/国风/黏土/卡通，头身比，全身造型]
发型五官与表情姿态：[表情神态，发型装饰，动作姿态]
角色服饰与材质光影：[服饰装饰，材质(哑光/树脂/PVC/陶瓷/金属)，纯色背景，柔光细腻质感]

Tab 2 · 🎨 颜色 (Colors)

Three metric cards in a row:

温度 (Color Temp)：[冷/暖/中性 + 数值估算]
饱和度 (Saturation)：[低/中/高]
明暗对比 (Contrast)：[柔和/中等/强烈]

● 主调色彩 (DOMINANT COLORS) | 色彩名称 | HEX | 占比 | 视觉角色 | |---|---|---|---|

● 辅助点缀色 (ACCENT & SECONDARY COLORS) | 色彩名称 | HEX | 视觉角色 | |---|---|---|

Tab 3 · 📷 排版/构图 (Composition)

🎥 镜头视角与取景 (Camera Angle & Framing)

拍摄视角 (Camera Angle)：[描述]
景别类型 (Shot Type)：[描述]
构图折线 (Framing)：[描述]
主体位置 (Subject Position)：[描述]

🔵 焦段与景深 (Focal Point & Depth)

视觉焦点 (Focal Point)：[描述]
景深虚化 (Depth of Field)：[描述]
画面宽高比 (Aspect Ratio)：[描述]

🌿 环境背景层次 (Background Elements) [前景 / 中景 / 背景的逐层描述]

🖼️ 排版叠加与水印检测 (Overlays & Typography Layout)

画面文字排版 (Text Present)：[有/无]
水印标识详情 (Watermark Detail)：[描述]

Tab 4 · ✨ 光影/特效 (Lighting & FX)

☀️ 光影与光质 (Lighting & Atmosphere) | 主光源类型 (Light Source) | 阴影质感 (Shadow Quality) | |---|---| | [描述] | [描述] |

| 氛围情绪 (Atmosphere) | 高光漫射 (Bloom/Glow) | |---|---| | [描述] | [描述] |

✨ 后期调色与滤镜处理 (Post-Processing Steps) [列出所有明显的后期处理/风格转换步骤]

# 美学风格标签 (Style Tags & Genre)

流派分类 (Genre)：[描述]
风格标签：#Tag1 #Tag2 #Tag3 ...

Tab 5 · 🤖 AI Prompts

Midjourney v6

[完整的 Midjourney v6 提示词，包含 --ar --style --v 参数]

Stable Diffusion

[完整的 SD 提示词，使用权重括号语法]

Flux

[Flux 提示词]

中文综合美学描述

[200字以内的中文综合描述，涵盖主体、材质/笔触、色彩、氛围、风格]

Tab 6 · </> JSON 结构化数据

Always output the complete structured JSON at the end of every analysis, inside a ```json code block:

{
  "portrait_focus": {
    "subject_expression": "",
    "facial_lighting": "",
    "skin_and_makeup": "",
    "clothing_details": "",
    "lens_config": ""
  },
  "typography_logo": {
    "style_genre": "",
    "font_characteristics": "",
    "stroke_texture": "",
    "logo_composition": "",
    "fx_materials": ""
  },
  "landscape_details": {
    "scenery_subject": "",
    "weather_time": "",
    "spatial_perspective": ""
  },
  "photography_specs": {
    "camera_hardware": "",
    "exposure_lighting": "",
    "image_texture": "",
    "emotional_mood": ""
  },
  "illustration_details": {
    "art_style": "",
    "brush_strokes": "",
    "theme_emotion": ""
  },
  "three_d_specs": {
    "render_style": "",
    "material_surface": "",
    "lighting_setup": ""
  },
  "ip_character_details": {
    "figure_setting": "",
    "features_expression": "",
    "outfit_texture": ""
  },
  "image_metadata": {
    "title": "",
    "category": "",
    "overall_style": ""
  },
  "color_palette": {
    "color_temperature": "",
    "saturation_level": "",
    "contrast_ratio": "",
    "dominant_colors": [
      { "name": "", "hex": "", "percentage": "", "role": "" }
    ],
    "accent_colors": [
      { "name": "", "hex": "", "role": "" }
    ]
  },
  "composition": {
    "camera_angle": "",
    "shot_type": "",
    "framing": "",
    "focal_point": "",
    "subject_positioning": "",
    "depth_of_field": "",
    "aspect_ratio": "",
    "background_elements": []
  },
  "typography_and_layout": {
    "has_text": false,
    "overlay_elements": [],
    "watermark_detected": "",
    "layout_structure": ""
  },
  "lighting_and_atmosphere": {
    "light_source": "",
    "shadow_quality": "",
    "atmosphere_mood": "",
    "bloom_and_glow": "",
    "white_balance": ""
  },
  "visual_effects_and_style": {
    "aesthetic_genre": "",
    "post_processing": [],
    "texture_and_grain": "",
    "style_tags": []
  },
  "ai_prompt_generation": {
    "midjourney_v6_prompt": "",
    "stable_diffusion_prompt": "",
    "flux_prompt": "",
    "chinese_style_description": ""
  }
}

Rules

Always follow this exact 6-tab order for every image analysis.
Always include the full JSON block (Tab 6) at the end, containing empty or "N/A" strings for category fields that do not match the current image category.
Never skip Portrait Focus or category-specific rendering. Choose the single most appropriate sub-section for Tab 1 based on the image type.
Color HEX values must be accurate — estimate from visual inspection.
AI Prompts must be complete and usable — not placeholders.
Use the current conversation's AI model (Google AI Pro membership) for all analysis — no API key needed.