I don't have an answer, but this reference can help you.

Preacher, K. J., Zyphur, M. J., & Zhang, Z. (2010). A general multilevel SEM framework for assessing multilevel mediation. Psychological Methods, 15(3), 209–33. https://doi.org/10.1037/a0020141

Good question! I have not had the same analytical experience so far, but I would share the way that I have in mind. Probably it works! Imagine you have A (predictor), B (mediator 1), C (mediator 2), and D (dependent V). You then have already several examinations for all possible direct effects (B on A; C on B; D on C and so on). Now you need to add this command:

MODEL INDIRECT:

D IND A;

This gives you all the possible mediation effects including one or two meditators

I hope it works!