| name | my-fetch-article |
| description | Use when you need to fetch and extract a web article's content as clean markdown — handles blogs, docs, PDFs, arxiv. Also use for "fetch this article", "read this URL", or "extract content from". |
| argument-hint | < URL > |
/my-fetch-article — Web Article Fetcher
Fetches web article content and converts to clean markdown. Handles blogs, documentation, PDFs, and arxiv papers.
Usage
/my-fetch-article https://example.com/blog/post — fetch and display- Called by research pipelines for non-GitHub web URLs
Fetch Strategy
Standard web pages
WebFetch "<url>" with prompt "Extract the full article content as clean markdown preserving all code blocks, images, headings, and lists. Remove navigation, ads, footers, and sidebars."
PDF files
If URL ends in .pdf or content-type is application/pdf:
- Use the Read tool if downloaded locally
- Otherwise WebFetch with prompt focused on text extraction
Arxiv papers
If URL matches arxiv.org:
- Fetch the abstract page, not the PDF
- Extract title, authors, abstract, key sections
Output
# Article Title
**Source:** <url>
**Author:** <if available>
**Date:** <if available>
<article content in clean markdown>
When called by /my-research-targets, return raw content for classification.
Gotchas
- Paywalled sites return partial content — check if the article seems truncated
- JavaScript-heavy sites (SPAs) may return empty content via WebFetch — try with a prompt that asks for visible text
- Very long articles (>5000 words) should include the full content — never truncate code blocks
- Some sites block automated fetching — try different user agents or xcancel as proxy
Rules
- Read-only — never submit forms, create accounts, or bypass paywalls
- Preserve all code blocks, headings, lists — these are the highest-value content
- Strip navigation, ads, footers, cookie banners — only article content
- Never truncate code blocks even in very long articles
