Violation of the second law of thermodynamics with thermal radiation

If we have two cubic cavities both with perfectly reflecting walls, one much large than the other, both with a blackbody in them, then when they are both in thermal equilibrium with their surroundings and are at the same temperature the much larger box will permit a range of frequencies that the smaller one won’t.
Then if we use an optical filter that only allows low wavelength light to pass, we could allow light, and so energy, to go from one box to the other and we’d have a net flow of energy between two systems that are both at thermal equilibrium and at the same temperature.

Does this not violate the second law of thermodynamics?

To me it sounds reasonable that this should be possible, however the argument that should allow one to generalise Planck’s law of thermal radiation, derived specifically for a cubic cavity with reflecting walls, to cavities of arbitrary shape and made out of arbitrary material relies on this being impossible:
https://en.wikipedia.org/wiki/Kirchhoff%27s_law_of_thermal_radiation

I’d be very happy to read your insights.