Hacking Applications on Paper
summary
Demonstrates concrete security bugs found by early threat modeling for applications: a cow-monitoring API that allows users to submit oversized JSON payloads which can flood Kafka and cause outages; email normalization rules that incorrectly merge distinct addresses leading to collisions; and how eventual consistency in systems can enable double-counting of balances. Recommends running 4–5 person threat-modeling sessions using simple data-flow diagrams (DFDs), the STRIDE framework, and frequent re-evaluations.
AI-written summary · about
chapters (12)
- 0:00Paper hacking introduction
- 3:35What threat modeling isn't
- 7:51Threat modeling definition
- 11:46Good threat modeling how-to
- 15:16Team setup and prep
- 18:54Scope and priorities
- 23:07Threat modeling diagram components
- 26:39Building the threat model diagram
- 30:52Threat brainstorming
- 36:38Prioritization and updates
- 40:53Identified threats examples
- 47:00Summary and Q&A
related talks
43mThreat modeling
46mThreat modeling is on the rise
29mThreat Modelling 101
01h 00mThreat Modeling for Secure Software Design
27mCommon Application Security Mistakes that Enable Automated Attacks
46mInsane in the Supply Chain: Threat modeling for...
47mThreat Modelling Stories From The Trenches
49mThreat Modeling Toolkit