So, I have a few suggestions. Are you open to using GNU libraries in 
ZSH?  If so, I was thinking how ZSH is so powerful (and easy to use), 
that it can actually handle all types of data -- like all 256 characters 
of ASCII.  So with multibyte turned off that means binary, even if its a 
little slow.  But here is the problem with dealing with binary, or 
unusual characters through ZSH, every operation requires using several 
ZSH builtin commands / variable expansions.  This makes a process that 
converts data into something more usable on the input or the output very 
CPU heavy and inefficient.  My suggestion is to add a few options to the 
sysread and syswrite builtins that come stock with ZSH in the system module:

  * -d : (sysread) to make the variable created an integer type
    containing the numerical value corresponding to the raw byte value,
    for -s 1; and for -s 2+ (reading more than a single byte), make it
    an array with each indice containing a string of decimal digit(s),
    whose numerical value is generated the same way, representing the
    value for each byte.
  * -h : (sysread) to make a string value that is two characters per
    input byte, representing the value of the byte in hexadecimal
    (padded with 0 if less than 16, and obviously as with above, a null
    would equal 00).  This would make Zsh able to function for whatever
    given needs (boot environments, less prevalent unix-like systems)
    without having to supply an xxd program, and it would be much more
    efficient processing when any math is involved in the script.

-- 
Micah micah@askmicah.net <mailto:micah@askmicah.net>
AskMicah.Net <http://askmicah.net>, Problem Solving Agency