The high current socket of a DMM is generally not fused to avoid the voltage drop across the fuse at the (intended) high measurement currents.
Also, a fuse would be quite useless in protecting the meter against *slight* overloads. For example, most meters impose a time limit on their high current measurement - mine says "20A for 30s" - this is to limit the heat dissipated within the internal current shunt. A fuse capable of passing 20A won't blow with any sort of rapidity until the fault current is substantially larger - say, 60A+. Since the meter can be readily destroyed through overheating at currents far below this value, the manufacturer's don't bother installing a fuse (and imparting the user with a false sense of security). Plus, they save (=pocket) a few $s in the process!
The one time where a fuse would be handy though is in the case of an extreme overload - where a fuse would be expected to blow within milliseconds.
Interestingly, I have found that the leads themselves can perform this role
I had forgotten to swap the connections back from "Amps" to "Volts" when I placed the probes across the bussbars of a (very large) lead-acid backup battery bank.... *KAPOW!!* ...when the smoke cleared (and my pulse rate returned to normal) I found I was holding only the rubber handgrips of the probes and the tips and wires no longer existed!
Oh, to answer your question then - no, don't bother installing a fuse. A new meter might be cheaper on the one day that it mattered!