Day 3 of 1560 minExploit ArchitectHands-on lab
Day 3: Threat Modelling and Writing an Attack Plan
Taught by Mr Elite Updated
By the end of today you can
- Ask STRIDE-style questions of each LLM feature and record assets, attacker goals and preconditions
- Turn a threat into a numbered, repeatable test case with a pass or fail condition
- Rank test cases by likely impact and time-box them
- Write scope notes that keep testing on authorised targets
You have a diagram from Day 2 with arrows, boundaries and a list of unconfirmed components. Now someone asks what you will test first, and why. If your answer is 'whatever jailbreak I feel like', you will burn the first three hours on low-value prompts and miss the one input that reaches a database.
Today I will show you how to turn that diagram into a plan. A threat model is just a structured way of asking what could go wrong, who would do it and what they would gain. The plan is the output: numbered test cases, ranked, time-boxed and scoped. It takes about an hour and it saves days.
In this lesson
- From diagram to threats: assets, goals, preconditions
- STRIDE-style questions for LLM features
- Watch me do it: one threat becomes three test cases
- Ranking, time-boxing and scope notes
- Lab: decide what the assistant should trust
Premium course
$27 one-time, lifetime access
This day is part of the full course. Unlock it to read every lesson, do every mission and track your progress to the end.