What problem does it solve? When the built-in WebFetch tool is refused with a 403 Forbidden, an empty body, or a "Just a moment…" bot-check page, you lose access to content that is often still publicly available. This Skill provides a ladder of fallback retrieval methods so the content can still be reached without guessing or fabricating answers. ## Core Features & Use Cases - curl with browser User-Agent retry: A bundled script fetches the URL, retries with a browser UA when plain curl is refused, converts the result to markdown, and prints the page headings for targeted reading. - Machine-readable endpoints: Falls back to context7 MCP for library docs, llms.txt files, raw.githubusercontent.com, JSON APIs, and RSS feeds that are built for programmatic reads. - Honest failure reporting: When every rung fails, the Skill instructs reporting the blocked source instead of silently answering from training memory. - Use Case: A documentation page on debezium.io returns 403 to WebFetch; run the fetch403.sh script to retrieve it as markdown, then read only the relevant sections by line range. ## Quick Start Use the fetch-403 skill to retrieve the content of the URL that WebFetch just refused with a 403 error.
