Scaling Privacy Threat Modeling: From Architects to Developers
summary
An open-source set of 12 privacy rules implements LINDDUN-derived checks as extensions to Threagile, taking a YAML architecture input and running a rule engine to flag concrete issues: data minimization failures, unauthorized disclosure, insecure storage, and re-identification risks. The authors find 61% of LINDDUN threats are codable, the codable set covers 76% of mapped incidents, and the chosen 12 rules detect about 67% of real-world incidents and 53% of LINDDUN categories. A demo flags unused sensitive fields (disability, ethnic origin) flowing to a browser.
AI-written summary · about
chapters (10)
- 0:00Introduction and agenda
- 0:56Privacy threat modeling
- 3:17Introduction to Privitect
- 4:58Architecture and framework
- 6:55Development methodology
- 10:13Implementation challenges
- 11:19Overview of privacy rules
- 11:54Sample rule implementation
- 14:02Working example and demo
- 16:04Conclusion and takeaways
related talks
44mA Taste of Privacy Threat Modeling
20mPrivacy Shift Left: A Machine-Assisted Threat Modeling Approach
01h 12mModelling Privacy Threats of Medical Data and Health Data Spaces
22mPrivacy Threat Modeling
06mData Privacy Concerns and Solutions When using LLMs in Production
01h 15mDATA PROTECTION ENGINEERING: WHAT IS THE ROAD AHEAD?
43mPanel: Privacy Design Patterns for AI Systems: Threats and Protections
37mFrom the Trenches Privacy Engineering Pitfalls Successes