redbook-cards-skill

将任意文章 HTML 页面或纯文本，自动切片生成小红书风格图文卡片组（封面 + 内容图 + 金句图），输出单个可直接在浏览器预览的 HTML 文件。触发词：小红书、图文卡片、切片、种草图、笔记卡片。

This skill appears coherent and limited to generating an HTML card set from provided article_content. Things to consider before installing/using: (1) The generated HTML links to Google Fonts (fonts.googleapis.com); opening the file in a browser will cause external network requests for fonts — replace with local fonts if you need offline/privacy guarantees. (2) The README/workflow suggests optionally fetching article HTML via an HTTP request and using Puppeteer/Playwright for screenshots — these are integration choices that can enable the agent to access remote URLs or run a headless browser; only enable those if you trust the environment and the URLs being fetched. (3) The skill sends the article text to whatever LLM you configure (this is required for content extraction), so avoid feeding sensitive/private content unless you trust the model/execution environment. (4) Review demo-output.html / EXAMPLE.md to confirm visual/style expectations before automating screenshots. No secrets or installers are required, and nothing in the skill suggests data exfiltration endpoints or hidden behaviors.

Type: OpenClaw Skill Name: redbook-cards-skill Version: 1.0.0 The 'redbook-cards-skill' bundle is a legitimate content transformation tool designed to convert articles into styled HTML image cards for social media. The skill instructions in SKILL.md and PROMPT.md focus entirely on content parsing, layout design, and CSS styling, with no evidence of data exfiltration, malicious execution, or prompt injection attacks. A provided Python utility in PROMPT.md for text cleaning uses the standard 'html.parser' library safely without any risky calls like 'eval' or 'os.system'.