docs: add AGENTS.md - #56
Conversation
Signed-off-by: YvesCesar <yvesamorim73@gmail.com>
vitormattos
left a comment
There was a problem hiding this comment.
I think AGENTS.md is useful here, but some parts seem too tied to the current implementation, with specific classes, methods and internal behavior.
My concern is that this may become outdated quickly after a refactor, or even make the agent treat current architectural limitations as rules that should be preserved.
I would keep AGENTS.md focused on stable operational guidance: project conventions, compatibility requirements, tests, commands and constraints. Implementation details should live in the project documentation, while desired architectural improvements are better tracked as issues.
It would also be useful to explicitly state that AGENTS.md should be updated whenever a valid change makes any of its guidance outdated.
| @@ -0,0 +1 @@ | |||
| @AGENTS.md | |||
There was a problem hiding this comment.
Isn't this file unnecessary?
The AGENTS.md is a pattern
There was a problem hiding this comment.
I think AGENTS.md is useful here, but some parts seem too tied to the current implementation, with specific classes, methods and internal behavior. My opinion is that this may become outdated quickly after a refactor, or even make the agent treat current architectural limitations as rules that should be preserved.
I would keep AGENTS.md focused on stable operational guidance: project conventions, compatibility requirements, tests patterns, commands or where can found the commands and constraints.
It would also be useful to explicitly state that AGENTS.md should be updated whenever a valid change makes any of its guidance outdated.
I added this documentation so that AI agents can understand the context of the project.
I haven't done a thorough review yet because I'm still getting my bearings here. I thought it would be more practical to open the PR for you all to review.