Description
PILLAR (Privacy risk Identification with LINDDUN and LLM Analysis Report) is a privacy threat modeling assistant that brings the power of large language models (LLMs) to the established LINDDUN framework. Simply feed PILLAR a natural-language system description and it will transform them into comprehensive threat models with limited input from users. It provides three variants of LINDDUN threat modeling, SIMPLE, LINDDUN GO (by simulating multi-agent collaboration in a virtual threat modeling workshop), and PRO.
Please see the repository for more details on PILLAR.
Related Publications
-
Majid Mollaeefar, Andrea Bissoli, Dimitri Van Landuyt, Silvio Ranise
PILLAR: LINDDUN Privacy Threat Modeling using LLMs
In: 2025 International Workshop on Privacy Engineering (IWPE'25) (news)
