RE: LeoThread 2025-03-29 04:21

Scaling Threat Modeling with AI: Generating 1000 Threat Models Using Gemini 2.0 and AI Security Analyzer

This blog post presents a method to produce high-quality threat models using Gemini 2.0 Flash Thinking and a series of prompts. The author found that the LLM performed poorly with prompts that were too long or too short but found that they could use multiple prompts that are in a “sweet spot” of length to science the desired results. They then used GitHub Actions to queue repositories into a pipeline that their AI Security Analyzer could then generate security design documentation, threat modeling, attack surface, and attack tree analysis for.

#technology #infosec #ai