The best situation is to organize abstractions into a subfolder and use [declare -path ./subfolder] to add it. That way you can have 1 main, obvious patch and everything else is more hidden.
A further step is provide a shell/batch script which then starts Pd with the main patch as a command line argument. With the new preferences system, you can also specify a config file to load as well. A good name for this script is something like "RUNME" :)
i also have seen this manner of working in other experienced users who send a folder with tons of .pd files (and it often takes a while to figure out which one to open). But, i agree this works and it's very convenient. I also think this needs some pd knowledge on the person receiving the folder, depending on the case. Especially if Its a patch from the extended era
As for making an "exporter", I believe Pd extended had something like this and it builds down to copying all basically Pd itself into a directory with the patches required. This could be accomplished with a GUI plugin using tcl or even just a bash script, modifying to core Pd GUI is not really required (hence the plugin mechanism). It's a shame not so many people are taking advantage of this without having to wait for someone to *possibly* add stuff to Pd itself.