I Built a Proof Gatekeeper That Can't Lie to Itself — and Found the One Trap It Can't See
zkVerify Agent Games Season 01 — Creator Track entry (#AgentGames) Everything below ran locally, on my own machine, against the tournament's official circuits. No mainnet transaction, no VFY spent, no live endpoint, no scored rounds played. I will say plainly where the local build ends and the real tournament begins — because that boundary is the most interesting thing I found. TL;DR I built the…
The Proof Gatekeeper track in zkVerify's Agent Games requires a decision engine that judges proof-backed payment requests and verifies each proof using Groth16. This system was built by the author to test the limits of zero-knowledge proofs. The author's local build successfully passed all test cases, with every decision correct and every provable proof verifying.
However, the designer discovered a trap - the PROOF_REUSED trap cannot be proven by the circuit, as the proof reference is not a circuit input. This means that enforcing reuse must be done off-circuit, by the platform's seen-set. The author emphasizes that a zero-knowledge proof does not make an agent honest, it only makes a specific claim checkable.
Rewritten by urgent.news; may contain errors — check the original.