Even if an algorithm to be incorporated in safety-critical software was doctored by rogue staff, why would?

Question

that matter ? Wouldn't it be fixed during "software integration" ? Isn't that a way of testing that your system is maintainable - injecting faults and assuring yourself that the faults are fixed ?

Answer 1

Somebody has to hold the keys to the vault.

If the trusted people act in a rogue manner then nothing is safe.

The choice of people for the critical position that controls what software is installed on critical computers, is the most important factor in ensuring safety-critical software is not corrupted.