Optimizing Agent Planning for Security and Autonomy

Indirect prompt injection attacks threaten AI agents that execute consequential actions, motivating deterministic system-level defenses. Such defenses can provably block unsafe actions by enforcing confidentiality and integrity policies, but currently appear costly: they reduce task completion rates and increase token usage compared to probabilistic defenses. We argue that existing evaluations miss a key benefit of system-level defenses: reduced reliance on human oversight. We introduce autonomy metrics to quantify this benefit: the fraction of consequential actions an agent can execute without human-in-the-loop (HITL) approval while preserving security. To increase autonomy, we design a security-aware agent that (i) introduces richer HITL interactions, and (ii) explicitly plans for both task progress and policy compliance. We implement this agent design atop an existing information-flow control defense against prompt injection and evaluate it on the AgentDojo and WASP benchmarks. Experiments show that this approach yields higher autonomy without sacrificing utility.

Key Contributions

• Autonomy metrics that quantify the fraction of consequential actions an agent can execute without human-in-the-loop approval while preserving security guarantees
• Prudentia: a security-aware agent design that combines richer HITL interactions with explicit planning for both task progress and policy compliance
• Empirical demonstration on AgentDojo and WASP that security-aware planning improves autonomy without sacrificing task utility under information-flow control defenses

🛡️ Threat Analysis

Details

Similar Papers