Down With Large Language Models!

Down With Large Language Models!

In the age of LLM infiltrating everything, even customer service, it’s easy to see it as a potential attack route – and a pretty straightforward one at that. LLM’s access goes way beyond what you might think. Think of using LLM Integration for an attack as similar to exploiting a server-side request forgery, going after parts of a system that are usually off-limits.

So, I recently had a fun time experimenting with prompt injection attacks in a PortSwigger lab. No beating around the bush – I kicked things off by asking the LLM about its connected APIs. Once I got a peek at those APIs, the possibilities started to open up. This particular web application was about to fall victim to an Excessive Agency Attack! While digging into the returned JSON data, I quickly found out that I had control over the user database. Oh, and there’s this funny little insider joke on TeamPrism – we call it “POOR CARLOS.” No idea what Carlos did to deserve the treatment, but it must have been quite a show.

Catch you later, hackers!

Bodega

2 Comments

  1. I was recommended this website by my cousin I am not sure whether this post is written by him as nobody else know such detailed about my difficulty You are wonderful Thanks

  2. Excellent blog here Also your website loads up very fast What web host are you using Can I get your affiliate link to your host I wish my web site loaded up as quickly as yours lol

Leave a Reply

Your email address will not be published. Required fields are marked *