Bad channels in mne_inverse_operator

Hi MNE users,

I am using mne_inverse_operator command. The bad channels from my
observations, eeg data only, are different from those of the noise
covariance matrix. Does anyone know how mne_inverse_operator work under
this situation? Does MNE uses only the good channels in both files? I want
to double check this is the case.

Thank you very much for your comments and Happy Holidays!!

Best,