Hi @erlend,
Can you confirm that there isn’t another rule that is writing over the accessToken object?