Research2026-05-12
When Prompts Become Payloads: A Framework for Mitigating SQL Injection Attacks in Large Language Model-Driven Applications
Source: Arxiv CS.AI
arXiv:2605.10176v1 Announce Type: cross Abstract: Natural language interfaces to structured databases are becoming increasingly common, largely due to advances in large language models (LLMs) that enable users to query data using conversational input rather than formal query languages such as SQL....
arxivpapersprompting