i'm working on the clipboard program as we speak, totally reworking the interface.
my inclination is to do i think as you suggested.
first, capture not in plaintext but in native format, ie rtf if available, images if available (user will be able to say only capture plaintext).
then you wil be able to use those clips (export/save/copy/paste) as rtf, html, plaintext.
im open to more specific ideas about how the user interface should facilitate this, but bottom line is the program will capture in those formats not just plaintext as it does now, so those will be things we can do.