>how would we know which mutations are broadly beneficial or not?

In general terms we know what a good specimen looks like and the specific traits that they have. We can only build to that because:

>sickle cell syndrome are also more resistant to malaria

This is evolution by process of elimination, and unfortunetaly elimination means lack of breeding from disadvantaged phenotypes. We've largely uncoupled reproduction from evolutionary advantages anyway, and just like humans are so dedicated to controlling the world in they inhabit, we'll insist on controlling the finest details of reproduction too.