I was going to try to fix these myself, but am having trouble getting the vignette to compile, so I didn't want to start editing math that I couldn't easily check.
I suggest modifying the p_E = C(p_i, w_i) equation so that it more clearly denotes p_i and w_i as vector inputs. to me, p_i and w_i read as a single prediction from and weight for model i, not as a vector of predictions and weights. So maybe adding some boldmath notation to indicate vectors would be helpful here.
We come back to using the notation C to refer to the combination function in section 3.2 but C was only demonstrated in section 2 as applying to point forecasts. I suggest updating the equations with probabilistic forecast combinations to include C, even if just as an intermediate step, like F_LOP(x) = C( \bm w, \fancy F) = \sum_i w_i F_i(x) where I'm using \bm w to indicate some vector representation of w and \fancy F to stand in for some kind of "fancy" math symbol that we could define to represent the array of probabilistic predictions.
I was going to try to fix these myself, but am having trouble getting the vignette to compile, so I didn't want to start editing math that I couldn't easily check.