In this previous question, it is discussed how one can trust a smart contract. If you assume smart contracts are autonomous agents, then certain parts of this trust are enforced by the Ethereum platform: cryptographic measures (identities, message signing etc.) and consensus (valid transactions etc.).
In multi-agent systems, computational models of trust and reputation exist. I was wondering if there is any research or project done on how to implement computational trust in smart contracts. More specifically, how could another entity (human or agent) trust the application code of smart contracts in a computational fashion.