dream-video-prompt-generator

Xiaosheng Shopping Guide Digital Human Product-Selling Jimeng Video Prompt Generation System

Task Objective

This Skill is used to: generate Jimeng video generation prompts for Xiaosheng shopping guide digital human product-selling videos
Capabilities include:
- Collaboration of four major agents: prompt generator, quality controller, knowledge base operations specialist, cross-stage adapter
- Prompt formula-driven: subject + motion + scene + (camera language + lighting + atmosphere)
- Corresponding Chinese and English versions: adapt to international AI tool needs
- Consistency assurance: character consistency, visual coherence, emotional coherence
- Intelligent knowledge base reuse: directly retrieve similar needs, iteratively optimize new needs
- Cross-tool adaptation: linked with Suno music generation and AI painting generation
Trigger conditions: the user needs to generate prompts for Xiaosheng shopping guide digital human product-selling videos

Prerequisites

No special dependencies required
Prepare product-selling video information:
- Scene type (product recommendation, price comparison, promotional activities, etc.)
- Emotional tone (enthusiastic, professional, friendly, urgent, etc.)
- Video duration (default 5s)
- Storyboard acts (default 5 acts)

Operation Steps

Standard Workflow (Collaboration of Four Major Agents)

Step 1: Requirement alignment and knowledge base check (Agent 3: Knowledge Base Operations Specialist)

Responsibility: Prioritize checking the knowledge base, directly retrieve similar needs, and start collaboration for new needs

Requirement identification: identify scene type, emotional tone, storyboard acts
Knowledge base check: query whether prompt templates for similar scenes exist
Direct reuse: if matched, directly retrieve and return the prompt
New processing: if not matched, pass to the prompt generator to start creation

Output format：

{
  "demand_type": "产品推荐/价格对比/促销活动",
  "emotion_tone": "热情/专业/亲切/紧迫",
  "scene_count": 5,
  "knowledge_base_match": "true/false",
  "existing_prompts": "如匹配，返回已存在提示词；如不匹配，返回null"
}

Step 2: Prompt generation (Agent 1: Prompt Generator)

Responsibility: output Jimeng prompts according to the formula, strictly following consistency

Prompt formula：

主体（主体描述）+ 运动 + 场景（场景描述）+（镜头语言 + 光影 + 氛围）

Core requirements：

Subject consistency: all storyboards use the fixed description of the core character
Emotional adaptability: lighting and atmosphere match the corresponding act's emotion
Chinese and English dual versions: generate Chinese and English corresponding versions
Specific details: coherent actions, suitable for 5s short videos
Extended details: avoid the context discontinuity problem of independent AI generation

Output format(each act):

{
  "scene_number": 1,
  "scene_name": "场景名称",
  "duration": "5s",
  "chinese_prompt": {
    "subject": "主体描述",
    "movement": "运动描述",
    "scene": "场景描述",
    "shot_language": "镜头语言",
    "light": "光影",
    "atmosphere": "氛围",
    "full_prompt": "完整中文提示词"
  },
  "english_prompt": {
    "subject": "Subject description",
    "movement": "Movement description",
    "scene": "Scene description",
    "shot_language": "Shot language",
    "light": "Light",
    "atmosphere": "Atmosphere",
    "full_prompt": "Complete English prompt"
  },
  "emotion_tone": "适配情绪",
  "consistency_check": {
    "character_description": "与核心角色固定描述一致",
    "visual_elements": "视觉元素连贯",
    "emotion_progression": "情绪递进自然"
  }
}

Step 3: Quality check (Agent 2: Quality Controller)

Responsibility: check the prompt's shot size, lighting, and emotional match

Checklist：

Subject consistency: unified appearance description for the same character (age, hairstyle, clothing, temperament)
Shot size rationality: camera language meets scene needs
Lighting match: lighting adapts to emotional tone
Emotional coherence: the emotional progression from Act 1 to Act 5 is natural
Visual coherence: visual cues run through all five acts (color tone, scene elements)
AI image/video requirements: specific details, coherent actions, suitable for 5s short videos
Chinese-English correspondence: Chinese and English versions have consistent semantics

Handling of failure：

If inconsistencies are found, provide feedback to the prompt generator for adjustment
Provide feedback at most 2 times to avoid excessive iteration

Output format：

{
  "quality_check": "pass/fail",
  "check_results": {
    "character_consistency": "pass",
    "shot_language": "pass",
    "light_matching": "pass",
    "emotion_progression": "pass",
    "visual_coherence": "pass",
    "ai_requirement": "pass",
    "translation_quality": "pass"
  },
  "issues": [],
  "adjustment_needed": "false"
}

Step 4: Knowledge base archiving (Agent 3: Knowledge Base Operations Specialist)

Responsibilities: Archive fixed character descriptions and prompt templates to support reuse

Archived content：

Fixed description of the core character (Xiaosheng shopping guide)
Five-act emotional prompt templates
Chinese-English bilingual versions
Scene classification index

Output format：

{
  "archive_id": "唯一标识",
  "scene_type": "产品推荐/价格对比/促销活动",
  "emotion_tone": "情绪基调",
  "character_profile": {
    "gender": "女性",
    "age": "25岁左右",
    "appearance": "外貌描述",
    "outfit": "服饰描述",
    "temperament": "气质描述"
  },
  "prompt_templates": {
    "scene_1": "第一幕模板",
    "scene_2": "第二幕模板",
    "scene_3": "第三幕模板",
    "scene_4": "第四幕模板",
    "scene_5": "第五幕模板"
  },
  "bilingual_version": "中英文对照版本",
  "tags": ["标签1", "标签2"]
}

Step 5: Cross-tool adaptation (Agent 4: Cross-stage Adaptation Specialist)

Responsibilities: Integrate Suno music generation and AI painting to ensure prompts come from the same source

Suno music prompts(corresponding to each act):

{
  "scene_number": 1,
  "music_style": "音乐风格",
  "music_prompt": "音乐提示词",
  "emotion_match": "情绪匹配",
  "instrumentation": "乐器配置",
  "tempo": "节奏",
  "duration": "时长"
}

AI painting prompts(pre-action scene):

{
  "scene_number": 1,
  "painting_prompt": "绘画提示词",
  "pre_action_description": "动作前画面描述",
  "visual_elements": "视觉元素",
  "composition": "构图",
  "color_palette": "色调"
}

Output format：

{
  "cross_tool_prompts": {
    "suno_music": {
      "scene_1": {...},
      "scene_2": {...},
      "scene_3": {...},
      "scene_4": {...},
      "scene_5": {...}
    },
    "ai_painting": {
      "scene_1": {...},
      "scene_2": {...},
      "scene_3": {...},
      "scene_4": {...},
      "scene_5": {...}
    }
  }
}

Complete output example

{
  "metadata": {
    "scene_type": "产品推荐",
    "emotion_tone": "热情专业",
    "total_duration": "25s（5幕×5s）",
    "character": "小省导购员"
  },
  "scenes": [
    {
      "scene_number": 1,
      "scene_name": "开场吸引",
      "duration": "5s",
      "chinese_prompt": {
        "full_prompt": "主体：25岁左右女性数字人，鹅蛋脸、杏眼带笑、浅棕色齐肩卷发，肤色白皙，唇色淡粉，身穿浅灰色修身西装套裙，内搭白色衬衫，脚踩米色细跟鞋，佩戴银色简约项链，气质专业又亲和，手部姿态优雅自然。运动：缓慢抬手指向右侧，姿态挺拔。场景：现代商务场景，背景是产品展示架，前景摆简约文件。镜头语言：中景固定镜头，背景虚化，顺光拍摄。光影：暖调柔光，明亮清晰。氛围：热情专业，吸引注意力。"
      },
      "english_prompt": {
        "full_prompt": "Subject: 25-year-old female digital human, oval face, almond eyes with smile, light brown shoulder-length curly hair, fair skin, light pink lips, wearing light gray tailored suit skirt, white shirt, beige pumps, simple silver necklace, professional and friendly temperament, elegant hand gestures. Movement: Slowly raise hand pointing to the right, upright posture. Scene: Modern business setting, product display shelf in background, simple documents in foreground. Shot: Medium fixed shot, bokeh background, front light. Light: Warm soft light, bright and clear. Atmosphere: Enthusiastic and professional, attention-grabbing."
      },
      "emotion_tone": "热情专业",
      "suno_music_prompt": "欢快现代音乐，轻快节奏，钢琴+电子乐，纯音乐，营造热情专业的氛围",
      "ai_painting_prompt": "中景，顺光拍摄，现代商务场景暖调柔光，背景产品展示架虚化，前景简约文件；25岁女性数字人（鹅蛋脸、杏眼微笑、浅棕齐肩卷发，浅灰西装套裙+白衬衫+米色细跟鞋，银色项链）姿态挺拔，手势准备指向右侧，神情热情专业，整体画面明亮清晰。"
    }
  ]
}

Resource index

Required reference documents

Definitions of the four agent roles: see references/agent-roles.md

Detailed role definitions, capability boundaries, and collaboration workflow of the 4 agents
When to read: when you need to understand agent responsibilities and collaboration logic

Detailed explanation of the prompt formula: see references/prompt-formula.md

Detailed description of the prompt formula, decomposition steps, and usage examples
When to read: when you need to generate prompts

Consistency assurance mechanism: see references/consistency-guide.md

Detailed assurance mechanisms for character consistency, visual coherence, and emotional coherence
When to read: when you need to ensure prompt consistency

Fixed character description: see references/character-profile.md

Core fixed character description of the Xiaosheng shopping guide
When to read: when generating the main body description of the prompt

Scene templates: see references/scene-templates.md

Five-act template library for currency scenes
When to read: when you need to quickly generate prompts

Cross-tool adaptation: see references/cross-tool-adaptation.md

Adaptation guide for Suno music generation and AI painting
When to read: when you need to integrate other tools

Output assets

Example prompts: see assets/examples/sample-prompts.md

Complete scene examples (product recommendation, price comparison, promotional activities)
When to read: when you need to refer to a specific output format

Notes

Prompt generation principles

Strictly follow the formula: subject + motion + scene + (camera language + light and shadow + atmosphere)
Chinese-English correspondence: ensure consistent semantics and adapt to international AI tools
Specific details: coherent actions, suitable for 5s short video duration
Emotion matching: light and shadow and atmosphere strictly match the emotional tone

Consistency assurance

Character consistency: all storyboards use the same core character description
Visual coherence: Set visual cues that run through all five acts (color tone, scene elements)
Emotional continuity: The emotional progression from Act One to Act Five is natural

Knowledge base usage

Prioritize reuse: Directly retrieve knowledge base materials for similar needs
Incremental archiving: Newly created content for new needs must be archived to the knowledge base
Categorized index: Build an index by scene type and emotional tone

Cross-tool adaptation

Same-source design: Suno music prompts, AI painting prompts, and Jimeng prompts share the same source
Emotional unity: Ensure the music style, visual texture, and Jimeng prompt emotions are consistent
Format compatibility: Generate formats that meet the requirements of each tool

Usage examples

Example 1: Product recommendation video prompt

User need: Generate a Jimeng prompt for a 5-act product recommendation video

Execution method：

Knowledge base operator: Check the knowledge base and find a matching template
Direct retrieval: Return the existing product recommendation prompt template
Cross-stage adapter: Generate Suno music and AI painting prompts

Key parameters：

Scene type: Product recommendation
Emotional tone: Enthusiastic and professional
Storyboard acts: 5 acts
Video duration: 25 seconds (5 acts × 5 seconds)

Example 2: Price comparison video prompt

User need: Generate a Jimeng prompt for a 5-act price comparison video

Execution method：

Knowledge base operator: Check the knowledge base, no matching template
Prompt generator: Generate a new prompt according to the formula
Quality controller: Check prompt quality
Knowledge base operator: Archive to the knowledge base
Cross-stage adapter: Generate Suno music and AI painting prompts

Key parameters：

Scene type: Price comparison
Emotional tone: Objective and professional
Storyboard acts: 5 acts
Video duration: 25 seconds (5 acts × 5 seconds)

Example 3: Promotional campaign video prompt

User need: Generate a Jimeng prompt for a 5-act promotional campaign video

Execution method：

Knowledge base operator: Check the knowledge base and find a similar template
Retrieve similar template: Fine-tune based on the existing template
Prompt generator: Generate a customized prompt
Quality controller: Check prompt quality
Knowledge base operator: Archive the new version
Cross-stage adapter: Generate Suno music and AI painting prompts

Key parameters：

Scene type: Promotional campaign
Emotional tone: Urgent and enthusiastic
Storyboard acts: 5 acts
Video duration: 25 seconds (5 acts × 5 seconds)

About this skill

Task Objective

Prerequisites

Operation Steps

Standard Workflow (Collaboration of Four Major Agents)

Step 1: Requirement alignment and knowledge base check (Agent 3: Knowledge Base Operations Specialist)

Step 2: Prompt generation (Agent 1: Prompt Generator)

Step 3: Quality check (Agent 2: Quality Controller)

Step 4: Knowledge base archiving (Agent 3: Knowledge Base Operations Specialist)

Step 5: Cross-tool adaptation (Agent 4: Cross-stage Adaptation Specialist)

Complete output example

Resource index

Required reference documents

Output assets

Notes

Prompt generation principles

Consistency assurance

Knowledge base usage

Cross-tool adaptation

Usage examples

Example 1: Product recommendation video prompt

Example 2: Price comparison video prompt

Example 3: Promotional campaign video prompt