You'll need some Linux experience, which will come to you as you play

Say your squeeze is at /tmp/squeeze and you're 'stuck' there in shell #1.
Start up shell #2 (telnet/ssh) and do
ls /tmp/squeeze. That is the root of
that filesystem. Now you just have to prepend /tmp/squeeze to anything you see in shell #1 to run and test whatever. Just be warned that libraries might not exist in the fw that are required for you run them.
ldd will tell you what libs are dependent and what you're missing, if anything. If you're not missing any libs then you can run from inside the fw itself and not in squeeze. If you're missing libs then you'll need to either add an extra path to /etc/ld.so.cache and rescan with
ldconfig -C /tmp/ld.so.cache, or copy the missing libs to /usr/lib. You might get conflicts just adding an extra path to /etc/ld.so.cache because the fw uses glibc-2.8 while squeeze uses glibc-2.10.
If you're wanting to use gdb or any system tools though, then theres no need really to leave the squeeze environment because /proc is mounted from the fw and therefore you have access to all pids from inside squeeze to attach to.
Is this clear enough? If you need to know anything else just say.
