One comment--this might have been caught if there were a .seq file being fprime-seqgen'd as part of CICD. Maybe consider adding that to check for future regressions like this?

I think it would be sufficient for the Ref integration test to set and save a parameter. Right now it's not doing that.

https://github.com/nasa/fprime/blob/devel/Ref/test/int/ref_integration_test.py

Ah, I was wrong, it does actually use seqgen as part of CICD. However, seqgen doesn't include the PRM SET commands. Maybe the easiest fix is to just add a param set cmd to the seq:
https://github.com/nasa/fprime/blob/devel/Ref/test/int/test_seq.seq

But your alternative also works, and seems like it addresses the issue more directly.