In a decentralized consensus network, AI agents can employ a Sybil resistance mechanism known as proof-of-identity via cryptographic puzzles. Each agent generates a unique signal, encrypted using a public-key cipher, which is then broadcast to the network. Agents verify the signal by solving a computational puzzle, with solutions forming a hierarchical terrace of validated identities. An attacker attempting to create multiple fake identities must expend computational resources proportional to the number of Sybils, making large-scale attacks economically infeasible.