RFC 7748 section 4.2 defines Curve448, then says it is birationally equivalent to an unnamed Edwards curve, then says that both of those curves are 4-isogenous to another Edwards curve, which it defines as edwards448/Goldilocks. python-rubenesque appears to use the parameters of the first unnamed Edwards curve instead of the actual edwards448 curve.
(Side nitpicks: "ed25519" and "ed448" are the names of signature algorithms, not curves, so should they really be aliases for edwards25519 and edwards448? And conversely, should "Goldilocks" be an alias for edwards448?)
RFC 7748 section 4.2 defines Curve448, then says it is birationally equivalent to an unnamed Edwards curve, then says that both of those curves are 4-isogenous to another Edwards curve, which it defines as edwards448/Goldilocks. python-rubenesque appears to use the parameters of the first unnamed Edwards curve instead of the actual edwards448 curve.
(Side nitpicks: "ed25519" and "ed448" are the names of signature algorithms, not curves, so should they really be aliases for edwards25519 and edwards448? And conversely, should "Goldilocks" be an alias for edwards448?)