Communitygithub.com

majiayu000/claude-skill-registry

Xiao Sheng Digital Salesperson Avatar Video Prompt Generation System. Based on four collaborative AI agents (Prompt Generator, Quality Controller, Knowledge Base Operator, Cross-Tool Adapter), it outputs bilingual (Chinese/English) prompts using the 'subject + movement + scene + (shot language + lighting + atmosphere)' formula for 5s short videos. Ensures character consistency, visual coherence, and emotional progression, with intelligent knowledge base reuse and cross-tool adaptation (Suno music, AI painting).

What is claude-skill-registry?

This skill provides a specialized system for generating video prompts for digital salesperson avatars, specifically tailored for short-form (5s) e-commerce videos. It leverages a multi-agent architecture comprising four distinct AI agents: a Prompt Generator that crafts prompts following a strict formula (subject + movement + scene + shot language + lighting + atmosphere), a Quality Controller that verifies consistency and emotional progression across multiple scenes, a Knowledge Base Operator that reuses existing templates for similar requests and archives new content, and a Cross-Tool Adapter that generates synchronized prompts for supplementary tools like Suno (music) and AI painting. The primary use case is for creators producing digital human e-commerce videos, ensuring character appearance, visual style, and emotional tone remain coherent across all scenes. The skill supports both Chinese and English prompt generation, features a structured JSON output for each scene, and enforces a five-step workflow that prioritizes knowledge base reuse before initiating new content generation. Complete reference documentation for agent roles, prompt formulas, consistency mechanisms, and scene templates are provided, making it a comprehensive solution for automated, high-quality video prompt creation.

Works with✓Claude Code~Codex CLI~Cursor
npx skills add https://github.com/majiayu000/claude-skill-registry/tree/HEAD/skills/ai-llm/dream-video-prompt-generator

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Xiao Sheng Digital Salesperson Avatar Jimeng Video Prompt Generation System

Task Objective

  • This Skill is used for: Generating Jimeng video prompts for Xiao Sheng Digital Salesperson avatar e-commerce videos.
  • Capabilities include:
    • Four-agent collaboration: Prompt Generator, Quality Controller, Knowledge Base Operator, Cross-Tool Adapter
    • Prompt formula-driven: Subject + Movement + Scene + (Shot Language + Lighting + Atmosphere)
    • Bilingual (Chinese/English) output: Supports international AI tools
    • Consistency assurance: Character consistency, visual coherence, emotional progression
    • Intelligent knowledge base reuse: Directly fetch for similar requests, iterative optimization for new requests
    • Cross-tool adaptation: Integrates Suno music generation and AI painting generation
  • Trigger Condition: User needs to generate prompts for Xiao Sheng Digital Salesperson avatar e-commerce videos.

Prerequisites

  • No special dependencies required
  • Prepare e-commerce video information:
    • Scene type (product recommendation, price comparison, promotional activity, etc.)
    • Emotional tone (enthusiastic, professional, friendly, urgent, etc.)
    • Video duration (default 5s)
    • Number of storyboard acts (default 5)

Operation Steps

Standard Workflow (Four-Agent Collaboration)

Step 1: Requirement Analysis & Knowledge Base Verification (Agent 3: Knowledge Base Operator)

Responsibility: Prioritize checking the knowledge base; directly fetch for similar requirements, initiate collaboration for new requirements.

  • Requirement identification: Identify scene type, emotional tone, number of storyboard acts.
  • Knowledge base verification: Check if prompt templates for the same scene type exist.
  • Direct reuse: If matched, directly fetch and return prompts.
  • New requirement processing: If not matched, pass to the Prompt Generator to start creation.

Output format:

{
  "demand_type": "product recommendation / price comparison / promotional activity",
  "emotion_tone": "enthusiastic / professional / friendly / urgent",
  "scene_count": 5,
  "knowledge_base_match": "true/false",
  "existing_prompts": "If matched, return existing prompts; if not matched, return null"
}

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

Responsibility: Output Jimeng prompts according to the formula, strictly adhering to consistency.

Prompt formula:

Subject (subject description) + Movement + Scene (scene description) + (Shot language + Lighting + Atmosphere)

Core requirements:

  • Subject consistency: All storyboard acts use the same core character description.
  • Emotional adaptability: Lighting and atmosphere match the emotion of each act.
  • Bilingual version: Generate Chinese and English versions.
  • Detailed specifics: Smooth action flow, suitable for 5s short videos.
  • Extended details: Avoid context gaps caused by independent AI generation.

Output format (per act):

{
  "scene_number": 1,
  "scene_name": "Scene name",
  "duration": "5s",
  "chinese_prompt": {
    "subject": "Subject description",
    "movement": "Movement description",
    "scene": "Scene description",
    "shot_language": "Shot language",
    "light": "Lighting",
    "atmosphere": "Atmosphere",
    "full_prompt": "Complete Chinese prompt"
  },
  "english_prompt": {
    "subject": "Subject description",
    "movement": "Movement description",
    "scene": "Scene description",
    "shot_language": "Shot language",
    "light": "Lighting",
    "atmosphere": "Atmosphere",
    "full_prompt": "Complete English prompt"
  },
  "emotion_tone": "Adapted emotion",
  "consistency_check": {
    "character_description": "Consistent with core character description",
    "visual_elements": "Coherent visual elements",
    "emotion_progression": "Natural emotion progression"
  }
}

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

Responsibility: Check the match of shot scale, lighting, and emotion in the prompts.

Checklist:

  • Subject consistency: Uniform appearance description of the same character (age, hairstyle, clothing, temperament).
  • Shot scale appropriateness: Shot language meets scene requirements.
  • Lighting match: Lighting matches emotional tone.
  • Emotional coherence: Natural emotion progression from Act 1 to Act 5.
  • Visual coherence: Visual cues run through all five acts (color palette, scene elements).
  • AI image/video requirements: Detailed specifics, smooth action, suitable for 5s short videos.
  • Chinese-English correspondence: Semantic consistency between Chinese and English versions.

Non-compliance handling:

  • If inconsistencies are found, provide feedback to the Prompt Generator for adjustment.
  • Maximum of 2 feedback rounds 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 Operator)

Responsibility: Archive core character descriptions, prompt templates to support reuse.

Archive contents:

  • Core character fixed description (Xiao Sheng salesperson)
  • Five-act emotional prompt template
  • Chinese-English bilingual version
  • Scene type index

Output format:

{
  "archive_id": "Unique identifier",
  "scene_type": "product recommendation / price comparison / promotional activity",
  "emotion_tone": "Emotional tone",
  "character_profile": {
    "gender": "female",
    "age": "around 25 years old",
    "appearance": "Appearance description",
    "outfit": "Outfit description",
    "temperament": "Temperament description"
  },
  "prompt_templates": {
    "scene_1": "Template for Act 1",
    "scene_2": "Template for Act 2",
    "scene_3": "Template for Act 3",
    "scene_4": "Template for Act 4",
    "scene_5": "Template for Act 5"
  },
  "bilingual_version": "Chinese-English bilingual version",
  "tags": ["tag1", "tag2"]
}

Step 5: Cross-Tool Adaptation (Agent 4: Cross-Tool Adapter)

Responsibility: Integrate Suno music generation and AI painting, ensuring prompts from the same source.

Suno music prompts (per act):

{
  "scene_number": 1,
  "music_style": "Music style",
  "music_prompt": "Music prompt",
  "emotion_match": "Emotion match",
  "instrumentation": "Instrument configuration",
  "tempo": "Tempo",
  "duration": "Duration"
}

AI painting prompts (pre-action frames):

{
  "scene_number": 1,
  "painting_prompt": "Painting prompt",
  "pre_action_description": "Pre-action frame description",
  "visual_elements": "Visual elements",
  "composition": "Composition",
  "color_palette": "Color palette"
}

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": "product recommendation",
    "emotion_tone": "enthusiastic and professional",
    "total_duration": "25s (5 acts × 5s)",
    "character": "Xiao Sheng salesperson"
  },
  "scenes": [
    {
      "scene_number": 1,
      "scene_name": "Opening Attraction",
      "duration": "5s",
      "chinese_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 language: Medium fixed shot, bokeh background, front light. Lighting: Warm soft light, bright and clear. Atmosphere: Enthusiastic and professional, attention-grabbing."
      },
      "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": "enthusiastic and professional",
      "suno_music_prompt": "Upbeat modern music, lively tempo, piano + electronic, instrumental, creating an enthusiastic and professional atmosphere",
      "ai_painting_prompt": "Medium shot, front light, modern business scene with warm soft light, bokeh product display shelf in background, simple documents in foreground; 25-year-old female digital human (oval face, almond eyes smiling, light brown shoulder-length curly hair, light gray suit skirt + white shirt + beige pumps, silver necklace), upright posture, hand gesture preparing to point right, enthusiastic and professional expression, overall bright and clear image."
    }
  ]
}

Resource Index

Essential Reference Documents

Four-Agent Role Definitions: See references/agent-roles.md

  • Detailed role definitions, capability boundaries, and collaboration flow of the 4 agents
  • When to read: When understanding agent responsibilities and collaboration logic is needed

Prompt Formula Details: See references/prompt-formula.md

  • Detailed explanation of the prompt formula, breakdown steps, usage examples
  • When to read: When generating prompts

Consistency Assurance Mechanism: See references/consistency-guide.md

  • Detailed assurance mechanisms for character consistency, visual coherence, emotional coherence
  • When to read: When ensuring prompt consistency is needed

Core Character Description: See references/character-profile.md

  • Core character fixed description of the Xiao Sheng salesperson
  • When to read: When generating prompt subject descriptions

Scene Templates: See references/scene-templates.md

  • Five-act template library for e-commerce scenarios
  • When to read: When quick prompt generation is needed

Cross-Tool Adaptation: See references/cross-tool-adaptation.md

  • Guide for Suno music generation and AI painting adaptation
  • When to read: When integrating other tools is needed

Output Assets

Example Prompts: See assets/examples/sample-prompts.md

  • Complete scene examples (product recommendation, price comparison, promotional activity)
  • When to read: When needing to reference specific output formats

Precautions

Prompt Generation Principles

  • Strictly follow the formula: Subject + Movement + Scene + (Shot Language + Lighting + Atmosphere)
  • Chinese-English correspondence: Ensure semantic consistency, suitable for international AI tools
  • Detailed specifics: Smooth action flow, suitable for 5s short video duration
  • Emotion matching: Lighting and atmosphere strictly match the emotional tone

Consistency Assurance

  • Character consistency: All storyboard acts use the same core character description
  • Visual coherence: Set visual cues that run through all five acts (color palette, scene elements)
  • Emotional coherence: Natural emotion progression from Act 1 to Act 5

Knowledge Base Usage

  • Prioritize reuse: Directly fetch existing knowledge base materials for similar requirements
  • Incremental archiving: New requirements must be archived into the knowledge base after creation
  • Category indexing: Index by scene type and emotional tone

Cross-Tool Adaptation

  • Same-source design: Suno music prompts and AI painting prompts originate from the same source as Jimeng prompts
  • Unified emotion: Ensure music style, image texture are consistent with the emotion of Jimeng prompts
  • Format compatibility: Generate formats that meet the requirements of each tool

Usage Examples

Example 1: Product Recommendation Video Prompts

User Requirement: Generate Jimeng prompts for a 5-act product recommendation video.

Execution method:

  1. Knowledge Base Operator: Check knowledge base, find a matching template.
  2. Direct fetch: Return existing product recommendation prompt template.
  3. Cross-Tool Adapter: Generate Suno music and AI painting prompts.

Key parameters:

  • Scene type: Product recommendation
  • Emotional tone: Enthusiastic and professional
  • Number of storyboard acts: 5
  • Video duration: 25 seconds (5 acts × 5 seconds)

Example 2: Price Comparison Video Prompts

User Requirement: Generate Jimeng prompts for a 5-act price comparison video.

Execution method:

  1. Knowledge Base Operator: Check knowledge base, no matching template.
  2. Prompt Generator: Generate new prompts according to formula.
  3. Quality Controller: Check prompt quality.
  4. Knowledge Base Operator: Archive into knowledge base.
  5. Cross-Tool Adapter: Generate Suno music and AI painting prompts.

Key parameters:

  • Scene type: Price comparison
  • Emotional tone: Objective and professional
  • Number of storyboard acts: 5
  • Video duration: 25 seconds (5 acts × 5 seconds)

Example 3: Promotional Activity Video Prompts

User Requirement: Generate Jimeng prompts for a 5-act promotional activity video.

Execution method:

  1. Knowledge Base Operator: Check knowledge base, find a similar template.
  2. Fetch similar template: Fine-tune based on existing template.
  3. Prompt Generator: Generate customized prompts.
  4. Quality Controller: Check prompt quality.
  5. Knowledge Base Operator: Archive new version.
  6. Cross-Tool Adapter: Generate Suno music and AI painting prompts.

Key parameters:

  • Scene type: Promotional activity
  • Emotional tone: Urgent and enthusiastic
  • Number of storyboard acts: 5
  • Video duration: 25 seconds (5 acts × 5 seconds)

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

majiayu000/claude-skill-registry

Builds, edit or validate Agent Skill through conversational discovery. Use when the user requests to "Create an Agent", "Optimize an Agent" or "Edit an Agent".

majiayu000/claude-skill-registry

Edit existing BMAD agents while maintaining compliance

majiayu000/claude-skill-registry

Edit existing BMAD agents while maintaining compliance

majiayu000/claude-skill-registry

Edit existing BMAD modules while maintaining coherence

majiayu000/claude-skill-registry

Edit existing BMAD modules while maintaining coherence

majiayu000/claude-skill-registry

Analyzes current state and user query to answer BMad questions or recommend the next workflow or agent. Use when user says what should I do next, what do I do now, or asks a question about BMad

majiayu000/claude-skill-registry

Build AI agents with Google ADK Python (Agent Development Kit). Use for multi-agent systems, workflow agents (sequential/parallel/loop), Vertex AI deployment, tool integration, human-in-the-loop.

majiayu000/claude-skill-registry

Build AI agents with Google ADK Python (Agent Development Kit). Use for multi-agent systems, workflow agents (sequential/parallel/loop), Vertex AI deployment, tool integration, human-in-the-loop.

majiayu000/claude-skill-registry

Portfolio allocation and rebalancing optimizer. Manages asset allocation across stocks/cash/bonds, performs periodic rebalancing, and ensures diversification according to market regime and risk tolerance.

majiayu000/claude-skill-registry

Create a hilarious and ultra-realistic video of an anthropomorphic animal acting like a human vlogger in a real-world setting.

majiayu000/claude-skill-registry

Speech-to-text transcription and translation via OpenAI Audio API -- models, response formats, timestamps, prompting, streaming, chunking, and diarization

majiayu000/claude-skill-registry

BlockRun works with Claude Code and Google Antigravity.

majiayu000/claude-skill-registry

This skill should be used when the user asks about libraries, frameworks, API references, or needs code examples. Activates for setup questions, code generation involving libraries, or mentions of specific frameworks like React, Vue, Next.js, Prisma, Supabase, etc.

majiayu000/claude-skill-registry

Upload and manage files using Google Gemini File API via scripts/. Use for uploading images, audio, video, PDFs, and other files for use with Gemini models. Supports file upload, status checking, and file management. Triggers on "upload file", "file API", "upload image", "upload PDF", "upload video", "file management".

majiayu000/claude-skill-registry

CRITICAL: Use for MolyKit AI chat toolkit. Triggers on: BotClient, OpenAI, SSE streaming, AI chat, molykit, PlatformSend, spawn(), ThreadToken, cross-platform async, Chat widget, Messages, PromptInput, Avatar, LLM

majiayu000/claude-skill-registry

AI and machine learning workflow covering LLM application development, RAG implementation, agent architecture, ML pipelines, and AI-powered features.

majiayu000/claude-skill-registry

Analyze images/audio/video with Gemini API (better vision than Claude). Generate images (Imagen 4), videos (Veo 3). Use for vision analysis, transcription, OCR, design extraction, multimodal AI.

majiayu000/claude-skill-registry

Analyze images/audio/video with Gemini API (better vision than Claude). Generate images (Imagen 4), videos (Veo 3). Use for vision analysis, transcription, OCR, design extraction, multimodal AI.

majiayu000/claude-skill-registry

Conduct domain and industry research. Use when the user says "lets create a research report on [domain or industry]

majiayu000/claude-skill-registry

Skill for discovering and researching autonomous AI agents, tools, and ecosystems using the AgentFolio directory.

Related Skills