WeChat Article Extractor

功能特性

安装

方式一:在 Claude Code / CodeX 中直接安装(推荐)

帮我安装这个 skill https://github.com/AngChow/wechat-article-extractor-skill.git

方式二:手动安装

git clone https://github.com/AngChow/wechat-article-extractor-skill.git
cd wechat-article-extractor-skill
npm install

使用方法

基本用法 - 从 URL 提取

const { extract } = require('./scripts/extract.js');
const result = await extract('https://mp.weixin.qq.com/s?__biz=...&mid=...&idx=...&sn=...');
if (result.done) {
 console.log('文章标题:', result.data.msg_title);
 console.log('公众号:', result.data.account_name);
 console.log('发布时间:', result.data.msg_publish_time_str);
} else {
 console.error('提取失败:', result.msg);
}

从 HTML 内容提取

const { extract } = require('./scripts/extract.js');
const html = await fetch(url).then(r => r.text());
const result = await extract(html, { url: sourceUrl });

高级选项

const result = await extract(url, {
 shouldReturnContent: true, // 返回 HTML 内容(默认:true)
 shouldReturnRawMeta: false, // 返回原始元数据(默认:false)
 shouldFollowTransferLink: true, // 自动跟随迁移后的公众号链接(默认:true)
 shouldExtractMpLinks: false, // 提取内嵌的微信公众号链接(默认:false)
 shouldExtractTags: false, // 提取文章标签(默认:false)
 shouldExtractRepostMeta: false // 提取转载来源信息(默认:false)
});

后处理:保存为 HTML 和 Markdown

const { convertToFixedHtml, convertToMarkdown } = require('./convert.js');
const fs = require('fs');
const path = require('path');
const outputDir = path.join(process.env.HOME, 'workspace_claude', 'wechat_articles');
fs.mkdirSync(outputDir, { recursive: true });
// 生成安全文件名(中文弯引号等特殊字符会被替换)
const dateStr = result.data.msg_publish_time_str
 ? result.data.msg_publish_time_str.replace(/[\/:]/g, '-').split(' ')[0]
 : 'unknown-date';
const safeName = (result.data.msg_title || 'untitled')
 .replace(/[\/\\:*?"<>|""'']/g, '_')
 .substring(0, 80);
const baseName = `${dateStr}_${safeName}`;
// 保存 HTML(data-src 已替换为 src,浏览器可直接打开查看图片)
fs.writeFileSync(
 path.join(outputDir, baseName + '.html'),
 convertToFixedHtml(result.data),
 'utf8'
);
// 保存 Markdown(含 YAML frontmatter)
fs.writeFileSync(
 path.join(outputDir, baseName + '.md'),
 convertToMarkdown(result.data),
 'utf8'
);

convertToFixedHtml(data)

convertToMarkdown(data)

响应格式

成功响应

{
 done: true,
 code: 0,
 data: {
 // 公众号信息
 account_name: "公众号名称",
 account_alias: "微信号",
 account_avatar: "头像URL",
 account_description: "功能介绍",
 account_id: "原始ID",
 account_biz: "biz参数",
 account_biz_number: 1234567890,
 account_qr_code: "二维码URL",
 // 文章信息
 msg_title: "文章标题",
 msg_desc: "文章摘要",
 msg_content: "HTML内容",
 msg_cover: "封面图URL",
 msg_author: "作者",
 msg_type: "post", // post|video|image|voice|text|repost
 msg_has_copyright: true,
 msg_publish_time: Date,
 msg_publish_time_str: "2024/01/15 10:30:00",
 // 链接参数
 msg_link: "文章链接",
 msg_source_url: "阅读原文链接",
 msg_sn: "sn参数",
 msg_mid: 1234567890,
 msg_idx: 1
 }
}

错误响应

{
 done: false,
 code: 1001,
 msg: "无法获取文章信息"
}

错误代码表

支持的文章类型

项目结构

wechat-article-extractor-skill/
├── scripts/
│ ├── extract.js # 核心提取逻辑
│ └── errors.js # 错误代码定义
├── convert.js # HTML 修复 + Markdown 转换模块
├── SKILL.md # Skill 定义文件(Claude Code / CodeX Skill 格式,含触发条件和后处理流程)
├── package.json # 项目配置
└── README.md # 本文件

依赖项

注意事项

许可证

贡献

更新日志

v1.1.0

v1.0.0