Zika virus (ZIKV) is a recently re-emerged flavivirus transmitted primarily through the bite of an infected mosquito, Aedes aegypti being the main vector. ZIKV infection is associated with a range of adverse effects; infection during pregnancy can lead to foetal abnormalities, including microcephaly. Lacking a licensed vaccine, or specific therapeutics, control of ZIKV transmission focuses on vector control. However, in most transmission settings, current methods are insufficient to successfully control ZIKV, or other similarly-transmitted arboviruses such as dengue and chikungunya viruses. This has stimulated interest in genetics-based methods, either to reduce the number of mosquitoes (population suppression), or to make mosquitoes less able to transmit (population modification). Here, we describe a method to selectively eliminate infected mosquitoes, using a virus sensor inserted into the mosquito genome and coupled to a quorum-counting lethal effector. In mosquitoes, ZIKV normally establishes persistent, lifelong infection; survival of these infected mosquitoes is crucial to transmission potential. Correspondingly, removal of infected mosquitoes can reduce vectorial capacity of a mosquito population, i.e. ability to transmit. Since relatively few mosquitoes become infected, typically <2%, engineered hypersensitivity to ZIKV would have only a modest population-level fitness cost, and lower still if transmission were successfully reduced by such means.