I wonder why you did not use the SRFI-4 names for your byte-vector
procedures.  Your rationale says

    "This SRFI is related to SRFI 4 (Homogeneous numeric vector
    datatypes), which also provides vectors of bytes. However, the
    extension described here does not require any extensions to the
    syntax of the underlying Scheme system."

so why not use exactly the same procedure names as SRFI-4, but simply
not support the external representation #u8(...), i.e. so that SRFI-4
is a pure extension of SRFI-66.  I just don't see any good reason to
invent a new SRFI-4 incompatible API for byte-vectors given that many
Scheme implementations currently support SRFI-4.

Marc