WeChat to Obsidian

将微信公众号文章剪藏到 Obsidian：使用真实无头浏览器加载页面，按原文顺序保留文字、图片和 Markdown 标题层级，图片自动下载到 attachments 目录。

This skill appears internally consistent with its purpose, but take these practical precautions before installing/using it: - Review and trust the agent-browser npm package before installing globally (check its npm page, repository, recent activity and any postinstall scripts). Installing global npm packages can add system binaries. - The skill will open web pages and run JavaScript in a headless browser context (agent-browser eval). That is necessary to extract content but means the browser may load third-party resources from the web. - The skill downloads images using curl with a WeChat Referer and will write files into the path you confirm. Always verify the suggested save path before confirming to avoid writing into sensitive directories. - Do not supply cookies, tokens, or private credentials to the skill. If your Obsidian vault contains sensitive files, consider running the skill against a copy or in a sandboxed environment first. - If you need stronger assurance, inspect the agent-browser package source or run it in an isolated environment (container/VM) before allowing it to write into your primary vault.

Type: OpenClaw Skill Name: wechat-to-obsidian Version: 1.0.3 The 'wechat-to-obsidian' skill is designed to scrape WeChat articles and save them as Obsidian notes. It uses 'agent-browser' to handle lazy-loaded content and 'curl' to download images with the required Referer headers. The instructions in SKILL.md include explicit safety steps, such as requiring user confirmation before file operations and filtering out non-content images like QR codes. No evidence of data exfiltration, unauthorized execution, or malicious intent was found.