Your goal: turn a detection idea into exact, tested code, and know what your rule can't see.
How this works (about 30 minutes)
- Marcus puts the question in the shared editor; pick Python or Java at the top.
- Think out loud. Ask about the log format and the rules before you code: what counts, which
window, which key.
- Say your approach and its time and memory before you write it.
- Run the public tests from the editor. Tell Marcus when you want him to look at your code.
- After the tests he asks what input breaks it, then changes what the attacker does.
- No scores during the call. Your report comes after.
What scores well
- Pinning down the window's edge and the key before coding
- One pass with a small state per key, and its cost said out loud
- Testing the edge of the window and other accounts' noise before Marcus asks
- Saying what your key misses, and how you'd catch tries spread across addresses
- A call on lockouts that weighs the users who get locked out
Tips
- Take Security Coach · Detection first.
- Work one small log by hand, including a failure exactly at the edge of the window.