Get Markdown

Fetch the markdown version of a URL to save context and reduce noise.

Usage

Run the script via Bash, passing the URL as the argument:

"${CLAUDE_PLUGIN_ROOT}/skills/get-markdown/scripts/get-markdown.sh" "$ARGUMENTS"

Present the script's stdout output to the user. On non-zero exit, display the stderr message prominently.

How It Works

The script applies the following logic in order:

1. Binary files: Rejects URLs pointing to non-text files (images, videos, archives, etc.).
2. Known URL patterns: Transforms documentation site URLs to their raw markdown equivalents and fetches via curl.
3. Plain text files: Fetches URLs with text file extensions directly — no transformation needed.
4. Tabstack fallback: Extracts markdown from all other URLs via the Tabstack API.

Pattern matching runs before the plain text extension check so that URLs like GitHub blob paths (which have text extensions but return HTML) are handled correctly. If a transformed URL returns HTML instead of markdown, the script automatically falls through to Tabstack.