The ubiquity of portable wireless-enabled computing and communications devices has stimulated the emergence of malicious codes (wireless worms) that are capable of spreading between spatially proximal devices. The potential exists for worms to be opportunistically transmitted between devices as they move around, so human mobility patterns will have an impact on epidemic spread. The scenario we address in this paper is proximity attacks from fleetingly in-contact wireless devices with short-range communication range, such as Bluetooth-enabled smart phones. An individual-based model of mobile devices is introduced and the effect of population characteristics and device behaviour on the outbreak dynamics is investigated. We show through extensive simulations that in the above scenario the resulting mass-action epidemic models remain applicable provided the contact rate is derived consistently from the underlying mobility model. The model gives useful analytical expressions against which more refined simulations of worm spread can be developed and tested.