Mobile agents can physically travel across a network, and perform tasks on machines, that provide agent hosting capability. These agents are autonomous; this property allows them to adapt themselves on a dynamic environment in a large network. Mobile agent programming is a complex task and it mixes business code and technical features. Because code development is costly, we use two kind of process algebra to write formal specifications. A first process algebra is used to the description of business part of the code. This formal language is an extension of pi calculus. A second process algebra allows us to architecture requirements. The second language is based on distributed pi calculus. We specify resource access and mobility schema. With two level specification, we drive more closely application building. The result is a layer application where architecture is a master feature and business part becomes its operator.