bawolff says it wouldn't be a problem to just link the curl docs.
Cogito says it would, since users would just ask Daniel for help, thinking xCurl was the same as curl, despite them having different implementations. This would place a large burden on Daniel.
Hopefully, by focusing on an example of how a support interaction could go (clueless user, lots of back and forth before the key issue is identified), it shows how this practice is good for neither users nor for the curl team who will end up dealing with the fallout.
I also address the MDN example by stating that I suspect the audience and reach of xCurl will create a larger support burden for curl than Mozilla would see in similar situations (aside: even though it wouldn't be reasonable in that case either!)
Unless it's done very carefully, pointing users to the parent project's documentation will cause a support burden on that project, and that's just not reasonable for Microsoft to do in this situation.