name: llms-docs-fetcher description: Discovers and downloads llms.txt and llms-full.txt from websites. Use when the user asks to "fetch documentation" or when starting an implementation task for a library where local documentation is missing or outdated. Trigger this skill to proactively search for llms.txt patterns on the project's dependency domains.

LLMS Docs Fetcher

This skill automates the discovery and local storage of llms.txt and llms-full.txt files.

Workflow

•Extract URL: Identify the base URL from the user request or project dependencies.
•Run Fetcher: Execute the fetcher script to check common locations.
•Storage: Files are saved to a docs/ directory with a GEMINI.md index.

Proactive Discovery

When working on implementation tasks:

•Missing Docs: If you are asked to implement a feature using a library (e.g., Panda CSS, Next.js) and no local documentation is found in docs/, proactively check if the library provides an llms.txt file at its root or /docs path.
•Outdated Docs: If local documentation seems outdated, re-run the fetcher to refresh the specific host directory.

Usage

Run the script using the following command:

bash


node scripts/fetch_llms_txt.cjs <URL1> [URL2] [URL3] ...

Example Requests

•
"Download the documentation from https://example.com and https://another-docs.com"
•
"Find the llms.txt for https://docs.github.com"
•
"Fetch docs for https://openai.com and https://anthropic.com"