The PN2222 is a big weak in driving already 1 x 2N3055. It may be OK with the current, but would likely not get rid of the heat. If a high current and thus more than 1 power transistor is planed one would also need a larger transistor (e.g. TO220 case, like D44H...) for the base current. At high current the current gain of the 2N3055 is only at some 20, so expect 1/20 the current and power in the driver.
The output stage as shown has quite some drop out voltage. One can get a little less drop with a Sislaki like configuration with the IC internal NPN, and external PNP and than NPN output transistors.
With the relatively low set voltage this may be an issue, but it depends on what voltage to start with.
The current limit at the base current is not very accurate, but the LM723 internal way is not accurate anyway. So it can be an OK compromise, saving on the shunt resistor power rating.
The feedback part of the circuit is not working right. That part is essentially all wrong: R1 should go to the output and the ref. signal has to go to the + input of the error amplifier.