Hi,
As the extended Hückel is just an initial guess, there is not need to worry about the gap at this point. You can proceed and carry out a single-point energy calculation with ridft (assuming you want to use DFT). After this, you can check the orbital energies with command "eiger". If everything is OK, you can proceed with the optimization. If you have gaps in the occupation (negative HOMO-LUMO gap), you could use the keyword $fermi to change the occupation.
Best,
Antti