Ignoring UNCONFIRMED bugs sometimes leads to missed regressions

Sometimes in UNCONFIRMED bugs bogus regression ranges can be found, e.g. https://bugzilla.mozilla.org/show_bug.cgi?id=1782411#c2, and so we decided to avoid setting status flags on UNCONFIRMED bugs (see #1627 and #1631).

This caused us to miss some regressions though, for example https://bugzilla.mozilla.org/show_bug.cgi?id=1789119 and https://bugzilla.mozilla.org/show_bug.cgi?id=1781167.
There was a needinfo to the regression author, so the bugs weren't completely out of the radar, but there was no follow-up from the author and the status flags are not set, which means the bugs were out of release managers and REOs radars during regression triage.

Can we improve the situation somehow? @gijsk any thoughts?
What about ignoring the UNCONFIRMED status and still update the flags in such cases:

• the "regressed by" field was set by an employee (or a person with editbugs privileges);
• the bug has a priority or severity;
• the "regressed by" was set for at least 4 weeks.

Another option would be to only set the nightly status to affected, so at least the bug is seen during regression triage.

No implementation files or tests are named. Start by reviewing issues #1627 and #1631 and the linked Bugzilla regressions to understand the current status-flag behavior; done means agreeing on and implementing a rule that keeps relevant UNCONFIRMED regressions visible during triage.