In the Storrs volume (Cornell-Silverman), Chapter 2 there is Lemma 1 stating that that if you have two semiabelian varieties over a normal scheme, then a homomorphism defined over an open dense extends uniquely over the whole base. The proof is not "idiot-friendly" (featuring phrases such as "one reduces immediately", "follows in a thousand other ways" not followed by a reference to EGA or whatever).
Can somebody give an idiot-friendly proof? Please do not give responses like "Go read EGA/SGA" without naming a specific theorem number, I have already been told that (and the question seems to be appropriate for MO, in my opinion).