Re: AGENTS.md

> From: "Dr. Arne Babenhauserheide" <[email protected]>
> Cc: [email protected],  [email protected],  [email protected],
>   [email protected],  [email protected]
> Date: Fri, 07 Aug 2026 14:28:47 +0200
> 
> Eli Zaretskii <[email protected]> writes:
> 
> >> From: "Dr. Arne Babenhauserheide" <[email protected]>
> >> Cc: Jean Louis <[email protected]>,  [email protected],  [email protected],
> >>   [email protected],  [email protected]
> >> Date: Fri, 07 Aug 2026 08:19:55 +0200
> >> 
> >> Eli Zaretskii <[email protected]> writes:
> >> 
> >> > Again, we created it to help users making innocent mistakes, using
> >> > whatever LLM-based agents, both proprietary and free.
> >> 
> >> Do the emacs modes that drive LLMs support AGENTS.md?
> >
> > There are no such modes.
> 
> I mean these, taken directly from package-list-packages:

They are not part of Emacs.  Almost all of them are from MELPA, which
package.el by default doesn't even access.  The Emacs project is not
responsible for those packages.

> If you don’t call these emacs modes or "that drive LLMs", please answer
> the question as if I had used the naming you use for them.

I have no idea which of these support AGENTS.md.

However, from my POV, these are not different from any other LLM-based
agent out there; the fact that they use Emacs infrastructure and Emacs
Lisp programs as a front end for LLMs is not relevant to this
discussion.  I don't think whether they do or don't has any special
importance for this discussion: they are just LLM harnesses, not
unlike any other harness.