Crowd Sourcing the Creation of Personae Non Gratae for Requirements-Phase Threat Modeling

Security threats should be identified in the early phases of a project so that design solutions can be explored and mitigating requirements specified. In this paper, we present a crowd-sourcing approach for creating Personae non Gratae (PnGs), which model attack goals and techniques of unwanted, potentially malicious users. We present a proof of concept study that takes a diverse collection of potentially redundant PnGs and merges them into a single set. Our approach combines machine learning techniques and visualization. It is illustrated and evaluated using a collection of PnGs collected from undergraduate students for a drone-based rescue scenario. Lessons learned from the proof of concept study are discussed and lay the foundations for future work.