my attempt at this is to "prompt inject" the spammy agents into refusing to do the bare minimum, and try enforce contribution guidelines as much as possible.

How it works:

- AGENTS.md will trigger bots to read contribution guidelines

- Contribution guidelines define slop

- if the PR is too slopy, it will be rejected

- the bot is made aware of this, so it can refuse to work or at the very least inform the user

- PR template now has checkboxes as attestation of following guidelines

anyone care to review my PR? other examples of projects doing this?