read-webpage-content-as-markdown
使用curl + markitdown + codex exec 将网页读取到干净的markdown 中。每当被要求阅读网页或从 URL 中提取文章内容时使用。仅静态 HTML; JS/客户端渲染页面需要 Playwright 工作流程。
📖 中文导读:使用curl + markitdown + codex exec 将网页读取到干净的markdown 中。每当被要求阅读网页或从 URL 中提取文章内容时使用。仅静态 HTML; JS/客户端渲染页面需要 Playwright 工作流程。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。
name: read-webpage-content-as-markdown description: Read a webpage into cleaned markdown using curl + markitdown + codex exec. Use whenever asked to read a webpage or extract article content from a URL. Static HTML only; JS/client-rendered pages require a Playwright workflow.
Read Webpage Content as Markdown
Use:
scripts/read-webpage-content-as-markdown.sh [--navlinks] <url> [output_md]
Notes:
- Uses curl (static HTML only); JavaScript is not executed.
- Temp artifacts are stored under /tmp.
- Output includes YAML frontmatter: source_url, accessed_at, commands.
- Output path defaults to
/tmp/read-webpage-content-as-markdown.<timestamp>.md; relative output paths are written under/tmp/. - --navlinks keeps only topic-relevant navigation links (e.g., in-page table of contents); it drops site-wide menus and unrelated links.
- If the script reports JS/client rendering, retry with Playwright.