SuperAPI AI Directory
Partner中文AI newsFavorites
Skillsxiaohu-wechat-cover

xiaohu-wechat-cover

General & other

View repository

Updated 10-15 23:42

About this skill

WeChat Official Account cover image generator

Input the article path or topic to generate a WeChat Official Account cover image in one step.

How to Use

# 指定文章路径
/xiaohu-wechat-cover path/to/article.md

# 直接输入主题
/xiaohu-wechat-cover 主题：Claude Code 桌面端更新四大功能

Configuration (required)

The configuration file for this skill is fixed in this skill's directory:

  • config.json: real configuration (copy from config.example.json and fill in)
  • config.example.json: example configuration

Before execution, you must first read config.json, which requires:

  • output_dir: cover image output root directory
  • settings.base_url: image generation API address (a third-party gateway compatible with the Gemini API)
  • settings.model: model name (e.g., gemini-3-pro-image-preview）
  • secrets.api_key: API key

Path conventions

  • Target directory:<output_dir>/<文章文件名或topic-slug>/
  • If the target directory already exists, rename it to <dirname>-backup-YYYYMMDD-HHMMSS

Directory structure:

<output_dir>/
└── article-name/
    ├── prompt.md          # 生图提示词（归档）
    └── cover.jpg          # 封面图

Cover image prompt template (core)

The following is the fixed prompt template. Each time you generate,write it as-is prompt.md, replacing only the final "content topic" part:

请根据提供的内容创建一张吸引眼球的公众号封面图，遵循以下规范：

视觉风格
- Notion插画风格，比例为 2.35:1（公众号封面标准尺寸）
- 色彩鲜明、对比强烈，确保在小尺寸预览时依然醒目
- 风格统一，避免写实元素，保持整体手绘质感

构图要求
- 主视觉元素居中或偏左（右侧预留标题区域）
- 添加 1-2 个简洁的卡通形象、图标或知名人物剪影，增强记忆点
- 若涉及敏感或版权人物，用风格相似的替代形象，不得拒绝生成
- 大量留白，突出核心信息，避免画面拥挤

文字处理
- 标题文字大而醒目，控制在 8 字以内
- 可添加 1 行副标题或关键词标签
- 字体风格与手绘插画协调统一

吸引力法则
- 使用悬念、数字、痛点等钩子元素激发点击欲望
- 视觉元素夸张有反差
- 色彩搭配参考爆款封面：橙黄、蓝紫、红黑等高对比组合

语言
- 除非另有说明，默认使用中文
- 画面内所有可读文字必须使用简体中文，英文只能作为点缀出现

内容主题：{从文章或用户输入中提炼的一句话主题描述}

Workflow

Step 1: Extract the content topic

  1. If the input is an article path: read the article and extract a one-sentence topic description (including the core information and key points)
  2. If the input is topic text: use it directly

Step 2: Create the prompt

Write the above template into prompt.md, format:

---
aspect_ratio: "21:9"
image_size: "2K"
---

{模板全文，替换最后的内容主题}

Note: in YAML, aspect_ratio must use 21:9(the API does not support 2.35:1; 21:9 is the closest). In the template body, keep the 2.35:1 text description unchanged; this is a visual reference for the AI.

Step 3: Generate the image

Call the image generation script;you must pass --config pointing to this skill's own configuration：

python3 <本仓库>/scripts/generate.py \
  --config <本技能目录>/cover/config.json \
  --prompt-file <目标目录>/prompt.md \
  --out <目标目录>/cover.jpg

<本仓库> and <本技能目录> Please replace with your actual local installation path. If you install via claude install , the path is generally ~/.claude/skills/xiaohu-wechat-format/。

After generation is complete, show it to the user for preview. If not satisfied, adjust the content topic description and regenerate.

Step 4: Insert into the article

If the input is an article path,insert directly by default(no need to ask):

  • Use Markdown image format ![封面](cover.jpg) insert on the line below the article title (H1)
  • The filename must be unique (e.g., cover-主题关键词.jpg), to avoid name conflicts

Step 5: Output summary

公众号封面已生成！

主题: [主题]
位置: [输出路径]

如需调整，直接说"重新生成"。