I’ve just published a Claude skill on Lawve AI called Mandatory Verification.
The idea is simple: AI should not be allowed to present non-trivial claims as true unless they have been properly checked.
So this skill pushes Claude to verify claims externally, prioritise real sources, check that information is up to date, and cite the basis for what it says.
I’d be very interested to hear how others are thinking about verification layers in AI workflows.