Implementing docs-hinting best practices for LLMs using Claude Code

Recently I implementing an LLM best practice for a docs site I wrote for my client Sensible using Claude Code. It’s an llms.txt file at the root of a docs GitHub repo that lists file directory structure and topics descriptions, to help LLMs interprete the docs site.

It was my first time using Claude Code, and I found it mind blowing to witness an LLM edit local files on my laptop. I’m now fired up about other docs-automation coding projects, and I’ve already made my first few commits on a project I’ll probably blog about soon!

To return to llms.txt – I’d heard that Claude Code could easily go off the rails without proper guidance, and I see it’s true! Even with such a simple task there were many ways I had to prevent it from barking up the wrong tree several times. The descriptions it generated in llms.txt needed some light hand editing too.

I also then used the llms.txt to implement another best practice (adding descriptions to the frontmatter of our MD topics, a long-neglected task). I’ll figure out how to automate updating llms.txt and keep the descriptions in sync using Claude Code and GitHub actions later.