I've now succeeded in generating the CTI corrections for 3 CCDs from the BOT run 13144. With Adam Snyder, we did the comparison below(CTI_Comparison.pdf) of the overscan analysis of CCD 11, comparing to Adam's original code. This seems to compare well, so that part of the code looks good.

I tried applying the correction, but that is blocked by DM-36280. However, I was able to apply the correction in a notebook, and I discovered two bugs in deferredCharge.py. At this point in flipData, the two dictionary calls to X_FLIP and Y_FLIP should use square brackets, not parentheses:
https://github.com/lsst/ip_isr/blob/564bce95634bcb13be2c8ab9beea4d036a4b12fd/python/lsst/ip/isr/deferredCharge.py#L627

Also, at this point:

https://github.com/lsst/ip_isr/blob/564bce95634bcb13be2c8ab9beea4d036a4b12fd/python/lsst/ip/isr/deferredCharge.py#L598

I believe that

image[amp.getBBox()].array[:, :] = correctedAmpData[:, :]

should read:

image[amp.getRawBBox()].array[:, :] = correctedAmpData[:, :]

After making these changes, the code ran and applied the CTI correction. The results on two amps are shown in the attached CTI_Test_3021120700215_28Sep22.pdf. I still don't think this correction is right, and am working with Adam to understand, but at least it is now running end-to-end.