@ansari_adin Good question, that’s what we are most concerned about. The trick is that we never touch the part that matters. Blop keeps two things separate: how a test finds content on the page (locators) and what the test expects to be true (assertions). Self-treatment only touches the first one. If a button has been moved or renamed, he can find it again. But if what you said should happen doesn’t happen, that’s failure, full stop. It never adopts any claim to beat the red test.
So a deliberate UI change usually looks like the locator wandering off, the button is still there, just in a new location. Regression looks like the expected result is not visible. We only auto-recover the first type, and even then any actual fixes come in as PRs which you review and merge yourself. Your exam cannot be rewritten behind your back. You get the trace and the difference and you make the call.
<a href