Day 10 of 1560 minExploit ArchitectHands-on lab
Day 10: Risk Scoring and Professional Reporting
Taught by Mr Elite Updated
By the end of today you can
- Score an AI finding with a CVSS vector plus written AI-specific reasoning
- Express likelihood for a probabilistic exploit as a measured success rate
- Write a report entry with reproduction steps and fix guidance a developer can act on
- Set tool rules in the lab that stop a hijack without breaking the agent's real job
You have spent nine days finding things. Some of them worked every time, some worked one run in five, and one of them was a system prompt that held nothing worth stealing. If you hand a developer a list that says High, High, High, they will either panic or ignore you. Both outcomes are your fault, not theirs.
Today is about the part of the job that decides whether anything gets fixed: scoring honestly, saying how often an exploit lands, and writing an entry someone can reproduce and act on. We also review the course so far, and the lab asks you to flip sides and write the defensive tool rules.
In this lesson
- Scoring: CVSS first, then what CVSS misses
- Likelihood when the exploit is probabilistic
- Watch me do it: from raw results to a report entry
- Writing the entry: summary versus detail
- Lab: set the tool rules from the other side
- Review: Days 1 to 9 in one pass
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.