@simonw - It's been 2.5 years with little progress finding

Edoardo Debenedetti 1,3*, Ilia Shumailov 2, Tianqi Fan1, Jamie Hayes 2, Nicholas Carlini 2, Daniel Fabian 1, Christoph Kern 1, Chongyang Shi 2, Andreas Terzis 2 and Florian Tramèr 3

1 Google, 2 Google DeepMind, 3 ETH Zurich

Large Language Models (LLMs) are increasingly deployed in agentic systems that interact with an external environment. However, LLM agents are vulnerable to prompt injection attacks when handling untrusted data. In this paper we propose CaMeL, a robust defense that creates a protective system layer around the LLM, securing it even when underlying models may be susceptible to attacks. To operate, CaMeL explicitly extracts the control and data flows from the (trusted) query; therefore, the untrusted data retrieved by the LLM can never impact the program flow. To further improve security, CaMeL relies on a notion of a capability to prevent the exfiltration of private data over unauthorized data flows.” loading=“lazy”>