Bootstrap AI guidance structure (AGENTS.md, agent agnostic skills)? #422

Currently I'm using Codex / Claude code to get things done, and as long as a single person does it I figure it can make sense to try help AI do a better job for everyone by accumulating some repo specific insights via AGENTS.md / CLAUDE.md.

In the sysand-index repo I have some AI maintained and AI consumed docs to help avoid needing to repeat myself guiding the AI. OK if I bootstrap something like that in this repo?

The key thing I want to ensure is that @andrius-puksta-sensmetry doesn't need to repeat himself correcting PRs made by me with AI, when I know there are patterns to document for the future to adhere to. I have previously written up some guidance in .gitignored files but i work from two computers and have done a git clean -xfd removing them once by mistake.