Or is that the secret sauce no one wants to share, the edge people see themselves having.
But some of the resellers have freebies, like:
https://app.primeintellect.ai/dashboard/environments?ex_sort...
Then for example you could write your own hook and convert existing code smell documentation which agents ignore into a format that works with the hook.
That's what I did; I did not use the library I linked to ~ It served only as an inspiration.
Instead, I let the agents create custom lint rules (using eslint, pylint, ...) and add custom coaching error messages based on where I want to take my codebase.
https://youtu.be/6AgndHSkHFI?t=238
(You could of course argue that you don't like the direction the rules push the agents in given in the example - some people don't prefer small functions everywhere - but that's not the point: The lint-hooks work in pushing the agent in the desired direction; If one desires something else they'd simply need different rules)
Just flagging big functions will make the agent write small functions ~ but not necessarliy in a good way
(for example the agent might just cut `doOneThingAndTheOther` in half and call the second half `doOneThingAndTheOther2`)