2. A mainstream publication republishes this because wikipedia is generally trustworthy
3. Someone else at wikipedia notices nonsense and removes article
4. Article gets reinstated using a referense to publication at (2)
It may sound weird but it has happened a few times. Wikipedia could probably do more to avoid circular references, but there's really no simple answer.
For an amusing but harmless example, see the Amelia Bedelia article. It's been long since reverted, but I'm sure there are many similar prank edits that survive to this day
1. make sure the primary source is publicly viewable and archiveable (be it a blog, acedemic profile, about us on a news site, or even a social media post).
2. ensure that there's some way to verify this as a primary source. This one pains me the most, but if we trust news sites to vet anonymous sources, may as well trust social media communities to vouch for a source really being that person. IME social media loves to call out lies, so them trusting a personality isn't as scary as it sounds.
3. be able to mark a source as "primary" to give a bit of distinction so we know there is some lack of vetting compared to a mainstream resource (which hopefully has such vetting built in). similar to [citation needed], it can be footnoted like [2, primary].
There is and always will be nonsense that penetrates through, but I don't think we should discount primary resources as much we do currently. That's not how real life reflects on such resources. Wikipedia has loosened up and started to use more social media as ciations, so this isn't a radical direction.
1. Someone adds obvious nonsense to wikipedia
2. A mainstream publication republishes this because wikipedia is generally trustworthy
3. Someone else at wikipedia notices nonsense and removes article
4. Article gets reinstated using a referense to publication at (2)
It may sound weird but it has happened a few times. Wikipedia could probably do more to avoid circular references, but there's really no simple answer.