feat: add llms txt

[ryanleecode]

No description provided.

[TorstenStueber]

@ryanleecode please add some context about what this will be used for? I mean, this is clearly to allow an LLM to understand the Host API, but is your intention then to create a skill (etc) from this?

[ryanleecode]

@ryanleecode please add some context about what this will be used for? I mean, this is clearly to allow an LLM to understand the Host API, but is your intention then to create a skill (etc) from this?

Could be anything but its just putting the docs in a machine readable format

[pgherveou]

Shouldn't we just stick to Markdown, and break them down into different linked page.?

llmstxt.org says:

We furthermore propose that pages on websites that have information that might be useful for LLMs to read provide a clean markdown version of those pages at the same URL as the original page, but with .md appended. (URLs without file names should append index.html.md instead.

Given the size of the json here, do you think this toon encoder / decoder is a wise choice?

I stumbled upon this blog post from hackmd https://homepage.hackmd.io/blog/2026/03/22/agentic-workflow-2026?utm_source=blog&utm_medium=featured-article

they use different accept header to serve different content. That's probably the best way to serve doc for llms. Base on the tools encoder / decoder that they are configured with they could query the doc in the most token efficient way

[pgherveou]

we should probably cut at words boundary rather than taking the first 100 characters

[pgherveou]

closing for now will check #19 that should be enough to work with AI agents