Is there AI auto debug with auto check-in?
We already have somewhat useful auto-complete and auto-generate code sections. The next step would be to have the "Run" command engages an Agent that can fix compile time bugs, link time bugs, and unit test failures. Then another Agent that auto-review for policy compliance, and finally auto-check-in into the nightly staging area. Does this exist already?
Current state of the art is not good enough, save for trivial changes. Cursor scans your whole codebase so it has more context than other models, but even that easily gets stuck in loops when the problem is a bit more involved, suggesting one change and then another back and forth, neither of which work.