Key agreement is a fundamental cryptographic primitive. It has been proved that key agreement protocols with security against computationally unbounded adversaries cannot exist in a setting where Alice and Bob do not have dependent variables and communication between them is fully public, or fully controlled by the adversary. In this paper we consider this problem when the adversary can "partially" control the channel. We motivate these adversaries by considering adversarial corruptions at the physical layer of communication, give a definition of adversaries that can "partially" eavesdrop and "partially" corrupt the communication. We formalize security and reliability of key agreement protocols, derive bounds on the rate of key agreement, and give constructions that achieve the bound. Our results show that it is possible to have secret key agreement as long as some of the communicated symbols remain private and unchanged by the adversary. We relate our results to the previous known results, and discuss future work.introduced wiretap model in which communicants use the noise in the channel to provide prefect secrecy for the communication without requiring a shared secret key. Security guarantee in both these models although strong, is only achievable under very restrictive assumptions. Perfect secrecy in Shannon's model requires communicants to share a secret key of the length at least the size of the message. Secure communication in wiretap setting is only possible if Eve's view of the codeword is "noisier" than the Bob's view of the codeword, which does not hold for settings that the eavesdropper is closer to Alice than Bob, and has a better communication channel to Alice. Thus positive result in both above approaches are under conditions that are of limited practical applications.Motivated by this observation, Maurer considered the more basic problem of secret key agreement (secure communication is possible, if one can have a shared key) where Alice and Bob want to share a secret key while the communication channel between them is eavesdropped by a computationally unlimited adversary. Maurer considered a minimum setting where Alice, Bob and Eve hold dependent variables X, Y and Z, with a joint distribution P XY Z , and Alice and Bob can interact over a public discussion (PD) channel: an authenticated channel that is fully visible by all system participants. The setting is "minimum" in the sense that it was proved [26] that without any initial joint distribution, secure key agreement is impossible. The joint partially leaked distribution P XY Z is the only resource of Alice and Bob, and so a basic question is when a secure shared key can be derived by interacting over the PD. The joint distribution can be generated by receiving transmission of a public random beacon (e.g broadcast by a satellite) that broadcasts samples of a random variable, and this is received by all parties over their individual channels. The distribution can also be simply generated by Alice sending a random string X n of length n to Bob, ov...