Maybe they can compile the bytecode to the x86 subset in this paper, and check if it is secure using their tool:

https://dl.acm.org/doi/pdf/10.1145/2254064.2254111