| name | webpage-reader |
| description | This skill should be used when the user wants to read, fetch, extract, or analyze the content of a web page from a URL. Use when the user provides a URL to access documentation, articles, blog posts, research papers, release notes, or any online content and needs clean, readable content without navigation clutter. |
| license | MIT |
Purpose
Extract readable content from a web page URL using the browser tool. Prefer the main article/docs body; drop nav, ads, cookie chrome, and boilerplate when summarizing or saving notes.
When to Use
Invoke this skill when:
- The user provides a URL and asks to "read", "fetch", "extract", or "get the content" of a page
- The user wants to analyze online documentation, API references, or technical guides
- The user shares a blog post, article, or research paper URL and needs the body content
- The user wants to save web content to a local Markdown file for offline reference or notes
- The agent needs to reference external content during a research or writing workflow
For multi-source research with citations, prefer deep-research. For PDFs or other binary formats, say the skill covers normal web pages and ask how the user wants to proceed.
Workflow
Step 1: Validate the Input
- Verify HTTP/HTTPS URL; if needed, ask for a valid URL.
- Raw
.md URLs → open with browser and treat as Markdown.
- Normal HTML pages → proceed with
browser.
- Multiple URLs → process sequentially; confirm count if many.
Step 2: Fetch with browser
Open the URL with browser and take the page content (snapshot / readable text / markdown as the tool provides).
- Prefer the main content region when the tool returns structure.
- If the tool returns noisy chrome, strip obvious nav/footer lines in your reply or saved notes.
- On failure (403, 404, timeout, login wall): report clearly; retry once on timeout.
Step 3: Deliver
- Default: return clean Markdown-oriented notes with the source URL at the top.
- If the user asked to save a file:
write (or edit) under the path they named.
- Optional: short summary + key excerpts when the page is long.
Step 4: Errors
| Situation | Response |
|---|
| 403 Forbidden | Site blocks automated access; try manual open or another source |
| 404 Not Found | Verify URL |
| Timeout | Retry once; then report |
| Login / paywall | Ask user to paste content or authenticate |
Remember
- Use
browser for page content.
- Ask the user in chat when choices are needed.
- For multi-source synthesis, hand off to
deep-research.
Example Usage
- "Read https://docs.example.com/api and summarize auth."
- "Fetch this blog post and save notes to
notes/article.md."
- "Extract the release notes from this URL."