You can flag it manually or you can delegate that job to your web of trust. Or some of each. And you’ll redelegate and route around bad actors if you think someone’s doing a bad job.