Nah, Im pulling from high assurance safety and security fields. Only approaches that worked for long periods basically create a design that enforces specific, good properties in every state. Other commenters are talking failed route of open ended design that tries to guess every form of malice. It's whitelisting vs AV signatures.
Explanations exist; they have existed for all time; there is always a well-known solution to every human problem — neat, plausible, and wrong.