My co-author and I are using gsem for our empirical analysis, but we have problems with how to interpret the results when we include interaction terms.
Our current model looks like this:

gsem (M1 <-c.X1##i.X2)(Y1 <- M1 c.X1##i.X2) logit nocapslatent

where X1 is our continuous independent variable and X2 is dichotomous. Since both of our dependent variables (i.e. M1 and Y1) are dichotomous, we use gsem with a link function of logit.

We are interested in the mediating effect of X1 on Y1 via M1, but we also hypothesize that this mediation effect is moderated by X2. We have struggled to demonstrate the substantive significance of the interaction effect. We use sembuilder to generate a graph, but it seems weird to show the result of an interaction term with a diagram. Also, we tried several margins commands, but it does not seem to reflect the indirect effect.

Do you happen to know any solutions? Thank you for your help!