Plotter's simulation model of Flare is only enabled when the user specifies a flare density value during analysis, which is applied to the curve's data points as I previously outlined. When exporting data, that flare simulation is always turned off so it is the raw EFS data either from 0.1 above FB+F or the Approximate CI method of speed point determination. So yes, ExpoDev always assumes that no flare model is in effect when it applies its flare compensation, because it knows that Plotter has exported the data without using its model of flare.
Neither Plotter nor ExpoDev present flare as stops, they both deal with flare as an average density that gets added, typically 0.02, which when translated to EFS adjustments in ExpoDev is a very small adjustment that is much, much less that 1 or even 1/2 a stop. It is the amount of exposure needed to reduce the negative density by that flare density amount. For instance with an EFS of 100, with a flare density factor of 0.02 added the EFS gets adjusted to only about 102. We're talking very small fractions of a stop.
Plotter's method of speed point determination just uses the standard model of speed point determination by (simulated) placing the ISO triangle for a curve with an average gradient of 0.62. Plotter interpolates a curve from the film data with a average gradient of 0.62 and then finds the speed point on that curve and assigns an EFS that the user stated as the ISO of the film (there's a little more to this that involves a calibrated personal reference point since not all films are the ISO they say they are and that has to be compensated for as well). Then Plotter finds the speed points for for each of the other curves and determines their EFS in relation to the found ISO point. It sounds to me that Plotter is using a different way of determining the EFS of the tested films from what you are doing.
However, back to what I think your question is: where does Plotter determine the exposure point, H, to be when it finds the adjustment to use for flare compensation. Plotter simply looks at the curve to find the FB+F density that equals the amount of flare specified, i.e. where the FB+F density is 0.02 typically. From that point it can then determine the amount of exposure needed to double the density from that section of the curve and thus knows what additional exposure amount will produce an increase in density of 0.02 for flare for that particular film and developer combination.