- Home
- Skills
- Alchaincyf
- Huashu Skills
- Huashu Xhs Image
huashu-xhs-image_skill
- Python
49
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill alchaincyf/huashu-skills --skill huashu-xhs-image- SKILL.md10.2 KB
Overview
This skill generates high-quality Xiaohongshu (Little Red Book) post images and cover art with a design-first workflow. It offers 2–3 clear design proposals, both AI-rendered and HTML-rendered fallbacks, then produces, previews, and packages images in a project-safe folder. It enforces vertical 3:4 layout, large readable typography, and strict no-watermark rules.
How this skill works
I analyze the provided note to extract theme, hero keywords, tone, and desired image count. I present 2–3 distinct design directions for the user to choose from. After selection, I build a detailed prompt and produce one AI image and multiple HTML screenshot variants when needed, then run visual checks and prepare files for upload.
When to use it
- When the user mentions terms like “小红书配图”, “小红书封面”, “小红书图片”, “做张小红书图”, or “笔记配图”.
- For single cover images, carousel sets, or informational image cards for Xiaohongshu posts.
- When you want a design-first process: proposal → selection → generation → preview → upload.
- When precise text rendering or tables must be 100% accurate (use HTML fallback).
- When you need consistent naming and folder organization for multi-variant outputs.
Best practices
- Always start with 2–3 concrete design proposals; do not skip the proposal step.
- Make the core keyword or number the visual hero (approx. 3x size, colored/emphasized).
- Keep main title ≤7 Chinese characters, subtitle ≤15 characters, and validate text rendering per image.
- Produce one AI result per chosen direction and HTML screenshots for text-accurate variants.
- Use vertical 1080×1440 px (3:4) and follow safe margins; avoid signatures or watermarks.
Example use cases
- Create a warm hand-drawn cover for a tutorial post highlighting a core number or tip.
- Design a dramatic dark-gold product announcement cover for a launch post.
- Produce an information-rich carousel where key stats are oversized and clearly ordered.
- Fallback to HTML screenshot when the AI misrenders Chinese text or when generating precise data tables.
- Generate several color/layout HTML variants alongside one AI-rendered image for client selection.
FAQ
If Chinese text renders incorrectly, I re-render that card using the HTML screenshot path so text is pixel-perfect.
How many design directions will you propose?
I propose 2–3 distinct directions, each with clear differences in style, color, and layout; the user picks one or requests a mix.
Will images include watermarks or author names?
No. I explicitly remove any personal signatures, watermarks, or author names in all outputs.