Exploiting LLM APIs with Excessive Agency

👩‍🎓👨‍🎓 Learn about Large Language Model (LLM) attacks! To solve this lab, we'll leverage a web-based LLM to delete the user carlos.

If you're struggling with the concepts covered in this lab, please review https://portswigger.net/web-security/... 🧠

🔗 Portswigger challenge: https://portswigger.net/web-security/...

🧑💻 Sign up and start hacking right now - https://go.intigriti.com/register

👾 Join our Discord - https://go.intigriti.com/discord

🎙️ This show is hosted by   / _cryptocat   ( ‪@_CryptoCat‬ ) &   / intigriti  

👕 Do you want some Intigriti Swag? Check out https://swag.intigriti.com

Overview:
0:00 Intro
0:46 Web LLM Attacks
1:36 What is a large language model?
2:12 LLM attacks and prompt injection
2:31 Detecting LLM vulnerabilities
2:50 Exploiting LLM APIs, functions, and plugins
3:09 How LLM APIs work
4:11 Mapping LLM API attack surface
4:55 Lab: Exploiting LLM APIs with excessive agency
5:11 Explore site functionality
5:35 Probe live chat feature
6:04 Attempt password reset
6:59 Interact with backend SQL API
7:49 Login as carlos (and delete the user)
8:05 Alternative approach (official)
8:39 Conclusion