agentleFS
Biraj2004Biraj2004/huashu-skills-english/huashu-image-upload-en/SKILL.md

huashu-image-upload

Skill3 starsChanged 6 months agoMITView on GitHub

One-click generation and upload of article illustrations to an image host, with automatic Markdown link insertion. Use when the user mentions "illustrations", "article images", "upload images", or "image for the article".

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 13 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈13

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After article proofreading is complete and illustrations are needed

    line 15
  2. Distinctive:

    When the article content is complete but images are missing

    line 16
  3. Distinctive:

    Auto-upload to image host: Call ImgBB to obtain permanent links

    line 25
  4. Distinctive:

    Verify display: Ensure all image links are valid

    line 27
  5. Distinctive:

    Maximum: No more than 10 (to avoid excessive interruption to reading flow)

    line 33
  6. Distinctive:

    Method: Use WebFetch to obtain Wikimedia Commons image links

    line 56
  7. Distinctive:

    Method: Call API to generate image, returns a temporary URL

    line 63
  8. Very specific:

    Use the project's existing /tools/upload_image.py script

    line 83
  9. Distinctive:

    Use permanent network links (not local paths)

    line 112
  10. Distinctive:

    Write a meaningful image description (alt text)

    line 113
  11. Distinctive:

    Keep descriptions concise and focused on the image content

    line 114
  12. Distinctive:

    Before executing an illustration task, confirm:

    line 144
  13. Distinctive:

    Read the full article, mark key sections

    line 160
  14. Distinctive:

    A: Not recommended. Local paths break when pasted into the WeChat editor — always use network links.

    line 217

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Image Generation & Upload

A professional capability for automatically generating illustrations for WeChat articles, uploading them to an image host, and inserting Markdown links.

When to Use This Skill

This skill is automatically loaded when the following scenarios are detected:

  • The user says "add illustrations now", "help me with illustrations", or "add images"
  • After article proofreading is complete and illustrations are needed
  • When the article content is complete but images are missing

Core Features

  1. Analyse illustration requirements: Determine 5–8 illustrations needed based on article content
  2. Source images from:
    • Public domain works (Wikimedia Commons, Google Arts & Culture)
    • AI-generated images (Volcano Engine doubao-seedream API)
    • Free stock libraries (Unsplash, Pexels)
  3. Auto-upload to image host: Call ImgBB to obtain permanent links
  4. Insert Markdown: Insert image links at the appropriate positions in the article
  5. Verify display: Ensure all image links are valid

Illustration Quantity Guidelines

  • Recommended: 5–8 images
  • Minimum: 3 images (cover + 2 body images)
  • Maximum: No more than 10 (to avoid excessive interruption to reading flow)

Illustration Placement Strategy

Mandatory placement:

  • ✅ Cover image (hero image): Below the title — always required
  • ✅ Core sections: 1 image per key section

Optional placement:

  • Theory support (e.g., referencing a classic work)
  • Data visualisation (comparison charts, graphs)
  • Case supplements (product screenshots, examples)

Image Source Priority

1️⃣ Public Domain Works (Highest Priority)

  • Sources: Wikimedia Commons, Google Arts & Culture
  • Suitable for: Classic artwork, historical portraits, ancient book covers
  • Advantages: Free, no copyright issues, high quality
  • Examples: Van Gogh's Starry Night, Kant portrait
  • Method: Use WebFetch to obtain Wikimedia Commons image links

2️⃣ AI-Generated (Recommended)

  • Source: Volcano Engine doubao-seedream-4-0-250828
  • Suitable for: Cover images, concept art, abstract themes
  • Advantages: Original, customisable, fast
  • Method: Call API to generate image, returns a temporary URL

3️⃣ Free Stock Libraries

  • Sources: Unsplash, Pexels, Pixabay
  • Suitable for: Cover images, concept art, background images
  • Advantages: Free, high quality, commercially usable (CC0)
  • Method: WebFetch search + download

4️⃣ Screenshots / Official Assets (Citation required)

  • Sources: YouTube, Bilibili, product websites, film posters
  • Suitable for: Case images (product screenshots, video stills)
  • Note: Cite the source; use under fair use principles
  • Method: Remind user to take their own screenshot, or generate an AI concept image as a substitute

Image Host Upload Workflow

Core Script

Use the project's existing /tools/upload_image.py script

Upload Method

# Call upload script (via Bash)
python3 /Users/alchain/Documents/writing/tools/upload_image.py <image URL or local path>

# Handle web images (e.g., AI-generated temporary links)
python3 /Users/alchain/Documents/writing/tools/upload_image.py "https://example.com/ai-generated.jpg"

# Handle local images
python3 /Users/alchain/Documents/writing/tools/upload_image.py "/Users/alchain/Pictures/image.png"

# Script automatically returns ImgBB permanent link

Fallback Mechanism

  • ✅ Prefer uploading to ImgBB image host (permanent links)
  • ⚠️ If upload fails, automatically use the original URL as a fallback
  • 📌 See /tools/README.md for image upload script documentation

Markdown Insert Format

![Image description](https://i.ibb.co/xxxxx/image.jpg)
  • Use permanent network links (not local paths)
  • Write a meaningful image description (alt text)
  • Keep descriptions concise and focused on the image content

Image Specifications

Size Specifications

  • Cover image: 1200 x 600px (16:9)
  • Body images: 800–1200px wide
  • Person/product: 800 x 800px (1:1)

File Specifications

  • Format: JPG (photos) / PNG (illustrations)
  • Size: < 500KB
  • Naming: lowercase, semantic, in English (e.g., mrbeast.jpg, starry-night.jpg)

Path Specifications

  • ✅ Network link: https://i.ibb.co/xxxxx/image.jpg (recommended, permanent)
  • ❌ Local path: No longer used (breaks when pasted into the WeChat editor)

Copyright Notes

  • ✅ Public domain works (author deceased 70+ years)
  • ✅ CC0-licensed images (Unsplash, etc.)
  • ✅ AI-generated images (confirm tool licence)
  • ⚠️ Screenshots/official assets require source attribution

Illustration Checklist

Before executing an illustration task, confirm:

  • Article has been proofread and content is finalised
  • Number of illustrations determined (5–8)
  • Illustration positions confirmed (cover + core sections)
  • All images obtained (AI-generated / public domain / stock)
  • All images uploaded to ImgBB image host
  • Image links are permanent network links (starting with https://)
  • Images inserted in Markdown (using network links)
  • Image descriptions (alt text) have been filled in
  • All image links verified as valid (accessible)

Execution Steps (5-Step Standard Workflow)

Step 1: Analyse article and determine illustration requirements

  • Read the full article, mark key sections
  • List illustration requirements (6–8 items)
  • Determine the position and purpose of each image

Step 2: Source / generate images

Based on source priority order:

  1. Public domain works → WebFetch to retrieve
  2. AI-generated → Call API
  3. Free stock → WebFetch search
  4. Screenshot assets → Remind user to provide

Step 3: Upload to image host, obtain permanent links

# Call upload script for each image
python3 /Users/alchain/Documents/writing/tools/upload_image.py <image URL>

Step 4: Insert images into article

![Image description](https://i.ibb.co/xxxxx/image.jpg)

Step 5: Verify display

# Check all image references
grep -n "!\[" "article-path.md"
# All image links should start with https://

Reference Resources

  • /tools/README.md — Image upload script documentation
  • /tools/upload_image.py — Image upload script
  • /writing/AI-image-generation-API.md — AI image generation API documentation
  • /writing/images/illustration-best-practices.md — Illustration best practices

Frequently Asked Questions

Q1: Will AI-generated image temporary links expire?

A: Yes. That's why you must call upload_image.py to upload to ImgBB and get a permanent link.

Q2: What if the image host upload fails?

A: The script has a fallback mechanism and will automatically use the original URL. However, it's recommended to check your ImgBB API configuration.

Q3: Does every article need illustrations?

A: Illustrations are strongly recommended, especially for long articles (3,000+ words). Short articles (under 1,000 words) may only need a cover image.

Q4: Can local image paths be used?

A: Not recommended. Local paths break when pasted into the WeChat editor — always use network links.

Technical Dependencies

  • Python script: /tools/upload_image.py (already in place)
  • API configuration: ImgBB API key (configured in ~/.zshrc)
  • AI generation API: Volcano Engine doubao-seedream (optional)
  • Tools: WebFetch (for retrieving public domain images)

Success Examples

  • "Why the Best Content Is Useless": 6 illustrations (2 public domain + 4 AI-generated)
  • "DeepSeek-OCR In-Depth Review": 5 illustrations, all uploaded to ImgBB, copy-pasted to WeChat without issue

Last Updated: 2025-11-07 Applicable Projects: WeChat article writing Maintained by: Huashu


By Huashu | AI Native Coder · Independent Developer WeChat Official Account "Huashu" | 300k+ followers | AI Tools & Productivity Flagship products: Kitty Catchlight (AppStore Paid Rankings Top 1) · Play DeepSeek with One Book

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.