AgentGuardian: Learning Access Control Policies to Govern AI Agent Behavior

TL;DR

AgentGuardian is a security framework that learns and enforces context-aware access control policies to ensure AI agents operate securely, accurately, and without malicious misuse, by monitoring behaviors and input patterns during a controlled learning phase.

Contribution

It introduces a novel adaptive policy learning framework for AI agents that enhances security and robustness through context-aware governance and control-flow analysis.

Findings

Effectively detects malicious or misleading inputs

Reduces hallucination-driven errors in AI agents

Maintains normal agent functionality during enforcement

Abstract

Artificial intelligence (AI) agents are increasingly used in a variety of domains to automate tasks, interact with users, and make decisions based on data inputs. Ensuring that AI agents perform only authorized actions and handle inputs appropriately is essential for maintaining system integrity and preventing misuse. In this study, we introduce the AgentGuardian, a novel security framework that governs and protects AI agent operations by enforcing context-aware access-control policies. During a controlled staging phase, the framework monitors execution traces to learn legitimate agent behaviors and input patterns. From this phase, it derives adaptive policies that regulate tool calls made by the agent, guided by both real-time input context and the control flow dependencies of multi-step agent actions. Evaluation across two real-world AI agent applications demonstrates that AgentGuardian effectively detects malicious or misleading inputs while preserving normal agent functionality. Moreover, its control-flow-based governance mechanism mitigates hallucination-driven errors and other orchestration-level malfunctions.