You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To help reviewers validate tests associated with patches, the agent could do things like:
check the tests fail (or don't make sense) before the patch;
mutate the code, check the tests fail;
given the bug description, generate potential breakages of an implementation (without showing the diff, so it's black-box), ask to verify if the tests cover them.
Activity
suhaibmujahid commentedon Oct 13, 2026
We could create a dedicated hackbot agent for this purpose.
marco-c commentedon Oct 13, 2026
I was thinking it would be part of the review agent, as an additional step of the review.