The critical first step in the clinical development of a malaria vaccine, based on live-attenuated Plasmodium falciparum sporozoites, is the guarantee of complete arrest in the liver. We report on an approach for assessing adequacy of attenuation of genetically attenuated sporozoites in vivo using the Plasmodium berghei model of malaria and P. falciparum sporozoites cultured in primary human hepatocytes. We show that two genetically attenuated sporozoite vaccine candidates, Δp52+p36 and Δfabb/f, are not adequately attenuated. Sporozoites infection of mice with both P. berghei candidates can result in blood infections. We also provide evidence that P. falciparum sporozoites of the leading vaccine candidate that is similarly attenuated through the deletion of the genes encoding the proteins P52 and P36, can develop into replicating liver stages. Therefore, we propose a minimal set of screening criteria to assess adequacy of sporozoite attenuation necessary before advancing into further clinical development and studies in humans.