We are facing a crisis of transportability in computational science. When proofs depend on massive toolchains, specific hardware environments, or "accepted consensus," trust stops traveling. It gets stuck locally. In cybersecurity, we solved a similar problem by moving to a Zero-Trust Architecture: never trust, always verify. Why haven't we done the same for high-stakes mathematical verification? In my latest research, P = NP, A Proof That Travels, I propose exactly this: a verification protocol where the "prover" is treated not as a respected colleague but as an untrusted adversary.