The empty cell in the panel/panellist performance is usually due to the fact that you do not have variability in your data (the assessor gives the same score to all the samples).
Indeed, most of the analysis requires dividing by the standard deviation or the variance, and both are equal to 0 when the scores are identical. And dividing by 0 is not really “appreciated” by computers…
Note that, at the panellist level, the agreement is measured through the correlation coefficient. And computing the correlation coefficient with two samples is quite critical, as you would need more samples!
More generally, running panel/panellist performance on dataset involving only two samples work from a mathematical point of view, but is not recommended as you would need more samples to get valuable results.