Abstract
The capability of LLM agents to function as the "brain" of a system fundamentally expands the scope of analysis beyond a standalone model. Consequently, safety is no longer only about input-output content alignment; it also concerns system behavior and real-world execution outcomes.
However, the current literature is fragmented across attack types, applications, and benchmarks, making it hard to explain why failures such as prompt injection, tool misuse, and memory poisoning often share the same structural cause and how they spread through an agent workflow.
In this survey, we treat isolation as a first-class principle for LLM-agent system safety. By isolation, we refer to the separation of user inputs, tool access, execution channels, inter-agent communication, and environment-originated context.
We organize the literature with a boundary-centric taxonomy of five boundaries: user-agent, agent-tool, agent-execution, agent-agent, and system-environment. This view helps identify where the loss of isolation first occurs, how compromise propagates across boundaries, and which defenses are most relevant at each interface.
We also summarize cross-boundary failure paths, discuss open challenges, and outline a research agenda for isolation-by-construction in future agent systems.
Blogger's Review: Isolation as a safety principle is crucial in LLM-agent systems, effectively addressing various security challenges. This research not only provides a framework for understanding system vulnerabilities but also lays a foundation for future security measures. By clarifying the roles and functions of different boundaries, researchers can better design defenses to enhance the overall safety of the system.