We should talk about: - input/output (is it trusted/untrusted?) - threats & mitigations
We should talk about: