I am not aware of any technical standard that would dictate what your assumptions should be for a case like this, but perhaps someone clever will chime in.
1.) The heat flux boundary condition isn't mysterious. You are specifying that you know the rate of heat transfer along that surface and that it is uniform. Whether or not the rate you specify and its uniformity match reality with respect to your actual application is a matter of engineering judgement.
2.) You should look at the ISO15099 standard to see how convection is approximated in the air cavity. It is a material model that approximates convection as diffusion by employing a pseudo-conductivity. It is a coarse assumption, but allows the problem to be solved by observing only the governing equation of heat conduction. If you want to capture the temperature field with a lot of fidelity a CFD model which solves the full set of fluid motion equations would be more realistic.
Dan