Dear Martijn,
You have probably already figured out that it works. But for future readers: For 2c ECP calculations, you can only use the dhf-type basis sets. You can use the standard auxiliary basis sets, i.e. without the -2c suffix. This is sufficient for the RI-J part and should also work rather well for RI-K. The same holds for relativistic all-electron calculations. Of course, tailored 2c cbas would be advantageous. So, we are working on that.
You can also confirm the results with analytical integrals or using a very large even-tempered reference basis set.
Best,
Yannick