Automated Threat Analysis and Management in a Continuous Integration Pipeline

Abstract

Security and privacy threat modeling is commonly applied to systematically identify and address design-level security and privacy concerns in the early stages of architecture and design. Identifying and resolving these threats should remain a continuous concern during the development lifecycle. Especially with contemporary agile development practices, a single-shot upfront analysis becomes quickly outdated. Despite it being explicitly recommended by experts, existing threat modeling approaches focus largely on early development phases and provide limited support during later implementation phases.

In this paper, we present an integrated threat analysis toolchain to support automated, continuous threat elicitation, assessment, and mitigation as part of a continuous integration pipeline in the GitLab DevOps platform. This type of automation allows for continuous attention to security and privacy threats during development at the level of individual commits, supports monitoring and managing the progress in addressing security and privacy threats over time, and enables more advanced and fine-grained analyses such as assessing the impact of proposed changes in different code branches or merge/pull requests by analyzing the changes to the threat model.