The question is not when to start threat modeling. It's when to stop
summary
Claim: threat models never finish, so stop when they become sufficient not complete. Practical stop rules include timeboxing (e.g. two hours per sprint), scoping by n=3 doomsday scenarios, or allocating 15 minutes per trust boundary. Tools used: draw.io + Excel + Word for diagrams, analysis and reports, local LLMs to generate mermaid diagrams, and Eraser-style diagramming; use recursive prompting to refine outputs. Prioritize fixes by time‑to‑fix and number of issues remediable, and reuse reference architectures and security patterns.
AI-written summary · about
chapters (13)
- 0:00Threat modeling basics
- 1:13When to start
- 4:01Threat modeling vs penetration testing
- 7:28Four question framework
- 10:21When good enough
- 13:32Threat modeling in SAM
- 17:29Start and end of threat modeling
- 21:24Practical ways to stop threat modeling
- 28:12Scope with doomsday scenarios
- 33:32Threat modeling sprints
- 35:21Security patterns
- 37:50Q&A tooling and AI
- 46:03Q&A prioritizing findings
related talks
49mThreat Modeling Panel
52mLessons From The Threat Modeling Trenches
43mThreat modeling
01h 00mThreat Modeling for Secure Software Design
58mThreat Modeling In And For Your Organization
57mUnlocking Threat Modeling
22mEmpowering Developers through Threat Modeling Excellence
57mIntroducing Threat Modelling To Established Teams