Unsure if I'll track here or in @isentinel/hooks but, add the ability to progressive overload only specific lint errors. Some errors are likely more fundamental than others.
For example, I ban beforeEach Jest hooks. I don't necessarily need to give an agent this context upfront, but it may be a good idea to surface this error at the point it happens (and maybe only once in a session(?)).
This should be a curated list though. We don't need to give all lint errors (e.g. formatting errors), but likely fundamental lints that do make a big difference. We could say "hooks are banned, prefer a factory function" in a PostEdit hook to tell the agent of this info. The primitive for this should live in this config - but need to figure out where best to surface it.
We will also need to curate the list of rules to go with it, and ensure turning off rules also turns off these checks.
Need to see if we want to also surface "clearly wrong code" vs. just surfacing detrimental coding standards.
Unsure if I'll track here or in @isentinel/hooks but, add the ability to progressive overload only specific lint errors. Some errors are likely more fundamental than others.
For example, I ban beforeEach Jest hooks. I don't necessarily need to give an agent this context upfront, but it may be a good idea to surface this error at the point it happens (and maybe only once in a session(?)).
This should be a curated list though. We don't need to give all lint errors (e.g. formatting errors), but likely fundamental lints that do make a big difference. We could say "hooks are banned, prefer a factory function" in a PostEdit hook to tell the agent of this info. The primitive for this should live in this config - but need to figure out where best to surface it.
We will also need to curate the list of rules to go with it, and ensure turning off rules also turns off these checks.
Need to see if we want to also surface "clearly wrong code" vs. just surfacing detrimental coding standards.