From Prompt Injections to SQL Injection Attacks: How Protected is Your LLM-Integrated Web Application?

Abstract

Demonstrates how prompt injection can be chained with traditional web attacks (SQL injection, XSS) in LLM-integrated applications.

Categories

Framework mappings

@inproceedings{pedro2024from,
  title = {{From Prompt Injections to SQL Injection Attacks: How Protected is Your LLM-Integrated Web Application?}},
  author = {Rodrigo Pedro and Daniel Castro and Paolo Molina and Nuno Santos},
  year = {2024},
  month = aug,
  booktitle = {USENIX Security 2024},
  eprint = {2308.01990},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2308.01990}
}