runwithdata

If something breaks, you likely want to be able to at least react and roll-back?

Question is, what is more effective? Using less tokens, writing cleaner code or, and I think that one matters most, number of inline comments on a PR the agent created that got actually fixed, ie defects. The time and token that go into the review/push cycle before merging. I personally would focus on the latter.

Makes a lot of sense to have the small improvements coming. So running /codex review again on the fixes is acting as a judge, correct? Wouldn’t it be worth considering flipping the reviewer again, so the fixes get an adversarial review again?