How would you design a process for threat modeling new features before code is written at HelloFresh?
