The objective of the study is to provide a way to delegate a proof of a property to a possibly untrusted agent and have a small certificate guaranteeing that the proof has been done by this (untrusted) agent. The key principle is to see a property as an encryption key and its proof as the related decryption key. The protocol then only consists of sending a nonce ciphered by the property. If the untrusted agent can prove the property then he has the corresponding proof term (λ-term) and is thus able to decrypt the nonce in clear. By sending it back, he proves that the property has been proven without showing the proof. Expected benefits include small certificates to be exchanged and the zero-knowledge proof schema which allows the proof term to remain secret. External agents can only check whether a proof exists without having any information about it. It can be of interest if the proof contains some critical information about the code structure for instance.