The fluences of gamma-ray bursts (GRBs) are measured with a variety of instruments in different detector energy ranges. A detailed comparison of the implied energy releases of the GRB sample requires then an accurate accounting of this diversity in fluence measurements that properly corrects for the red-shifting of GRB spectra. Here we develop a methodology to "k-correct" the implied prompt energy release of a GRB to a fixed comoving bandpass. This allows us to homogenize the prompt energy release of 17 cosmological GRBs (using published redshifts, fluences, and spectra) to two common comoving bandpasses: 20-2000 keV and 0.1 keV-10 MeV ("bolometric"). While the overall distribution of GRB energy releases does not change significantly by using a k-correction, we show that uncorrected energy estimates systematically undercount the bolometric energy by similar to5% to 600%, depending on the particular GRB. We find that the median bolometric isotropic equivalent prompt energy release is 2.19 x 10(53) ergs, with an rms scatter of 0.80 dex. The typical estimated uncertainty on a given k-corrected energy measurement is similar to 20%.