I don't really understand what you are trying to do, and I don't know anything
about JPF except what 10 seconds skimming the website tells me.  But I'm not
sure if you'll be able to do what you want (whatever it is).  The only way to
run code in another JVM is to start another process, and then it's just like
starting another process that was written in C++, Delphi, or Lisp -- there is
no "special" communication with it.  So, although you can certainly start
another Java process, there won't be the intimate connection between the pieces
that constitute a pluggable architecture.

You might be able to get somewhere by using RMI, or something similar, to
communicate between the two processes.  But that could be quite awkward unless
JPF can accommodate the extra failure modes that remote method invocation
brings with it (e.g. the network might go down).  Possibly you could define a
layer of your own generic wrappers/proxies which are all that JPF "sees", and
which use RMI (and its own proxies) internally without letting JPF know about
them.

You might also consider changing your architecture entirely so that the
framework "knows" about (and perhaps is built on) RMI.  You might even switch
to a different kind of component architecture like JavaSpaces (or whatever that
is called these days).

   -- chris