Xiaohongshu-Style Image Generation

Core Features

• Intelligently break down articles into multiple images (2-10 images)
• Automatically generate image prompts that match Xiaohongshu style
• Support multiple input sources (URL, file path, text content)
• Automatically create structured working directories
• Support post-processing (merge into long image, PPT, PDF)

Usage Scenarios

• "Generate Xiaohongshu images"
• "Create Xiaohongshu series images"
• "Turn this article into Xiaohongshu images"
• "Help me make Xiaohongshu-style graphics and text"

• Users only ask how to create (provide suggestions only)
• Users need a single cover image (use pw-cover-image)
• Users need images in other styles (use general image generation tools)

Usage Methods

# Generate from URL
/pw-redbook-image https://example.com/article

# Generate from text content
/pw-redbook-image "Article content..."

# Generate from markdown file
/pw-redbook-image path/to/article.md

Parameter Description

• Input Source (Required): Supports three formats
  • URL: Web address starting with http:// or https://
  • File path: Relative or absolute path, supports text files like .md, .txt
  • Text content: Directly passed article content string

Xiaohongshu Style Characteristics

• Ratio: Vertical (3:4 or 9:16)
• Style: Cartoon style, hand-drawn style
• Color Scheme: Morandi colors, cream white, off-white, light pink, mint green
• Text: Hand-drawn style text, prominent large titles, highlighted with highlighter lines
• Decorations: Cartoon elements, emoji icons, hand-drawn stickers, speech bubbles
• Layout: Streamlined information, ample white space, key points presented in bullet points

Content Breakdown Principles

• Cover Image: Strong visual impact, includes core title
• Content Image: Each focuses on 1 core viewpoint
• Ending Image: Summary/call to action/golden quote

• Simple viewpoints: 2-3 images
• Medium complexity: 4-6 images
• In-depth content: 7-10 images

File Management

Output Directory

redbook-{topic-slug}/
├── source.md              # Source article
├── prompts/               # Prompt files
│   ├── 01_封面图.md
│   ├── 02_内容图_核心观点1.md
│   ├── 03_内容图_核心观点2.md
│   └── 04_结尾图.md
└── images/                # Generated images
    ├── 01_封面图.png
    ├── 02_内容图_核心观点1.png
    ├── 03_内容图_核心观点2.png
    └── 04_结尾图.png

1. Extract topic from article title/content (2-4 words, kebab-case)
2. Example: "如何提升工作效率" →

   improve-work-efficiency

Conflict Resolution

redbook-{topic-slug}/

• Append timestamp:

  {topic-slug}-YYYYMMDD-HHMMSS

• Example:

  improve-work-efficiency

  already exists →

  improve-work-efficiency-20260123-143052

Source Files

source.md

source-{original filename}

Workflow

Step 1: Obtain Article Content

• URL: Use WebFetch to crawl webpage content
• File path: Use Read to read file content
• Text: Directly use the passed content

• Create working directory

  redbook-{topic-slug}/

• Save source file to

  source.md

  or

  source-{original filename}

Step 2: Analyze and Break Down Article

${SKILL_DIR}/references/文章拆解模板.md

• Extract topic and core viewpoints
• Determine image quantity (2-10 images)
• Generate breakdown plan, clarify core content of each image

• Simple viewpoints: 2-3 images
• Medium complexity: 4-6 images
• In-depth content: 7-10 images

Step 3: Generate Prompt Files

redbook-{topic-slug}/prompts/

• Cover image:

  ${SKILL_DIR}/references/封面图模板.md

• Content image:

  ${SKILL_DIR}/references/内容图模板.md

• Ending image:

  ${SKILL_DIR}/references/结尾图模板.md

• Naming format:

  01_封面图.md

  ,

  02_内容图_关键词.md

  ,

  03_结尾图.md

• Prompts use English, avoid Markdown format
• Include style keywords and color schemes

Step 4: Generate Images

• Prioritize using pw-danger-gemini-web
• If multiple skills are available, ask user to choose

• Output directory:

  redbook-{topic-slug}/images/

• Display generation progress
• Images sorted by serial number
• After generating the first image, confirm style before batch generating the rest

Step 5: Post-Processing (Optional)

brew install imagemagick
node ~/.claude/skills/pw-image-generation/scripts/merge-to-long-image.js \
  redbook-{topic-slug}/images \
  redbook-{topic-slug}/长图.png

node ~/.claude/skills/pw-image-generation/scripts/merge-to-pptx.js \
  redbook-{topic-slug}/images \
  redbook-{topic-slug}/配图.pptx

node ~/.claude/skills/pw-image-generation/scripts/merge-to-pdf.js \
  redbook-{topic-slug}/images \
  redbook-{topic-slug}/配图.pdf

Step 6: Output Summary

Xiaohongshu series images generated!

Topic: [Topic]
Number of images: [Quantity] images
Working directory: redbook-{topic-slug}/

Next steps:
- Preview all images to confirm style consistency
- If needed, use merge tools to generate long image, PPT or PDF

Best Practices

• Confirm style after generating the first image, then batch generate the rest
• Use the same style keywords and color scheme for all images
• Uniformly place author information at the bottom right corner, keep format consistent

• Each image focuses on one core viewpoint, avoid overloading content
• Ample white space to improve readability
• Titles should be large and prominent, highlight keywords with highlighter lines

• Use English prompts for better results
• Avoid Markdown format, use plain text description
• Clearly specify ratio (3:4 or 9:16), style (cartoon/hand-drawn), color scheme (Morandi colors)

Common Issues

improve-work-efficiency

improve-work-efficiency-20260123-143052

Template References

${SKILL_DIR}/references/

${SKILL_DIR}/references/

${SKILL_DIR}/references/

${SKILL_DIR}/references/

Extended Configuration

1. .pw-skills/pw-redbook-image/EXTEND.md

   (project-level)

2. ~/.pw-skills/pw-redbook-image/EXTEND.md

   (user-level)

• Default image quantity range
• Style keywords and color schemes
• Author information format
• Output directory naming rules