RePolicy: Reinforcement Learning for Safety-Policy Invocation in Agent Safeguards

2026-08-25

AI Summary

该研究针对语言模型代理的安全保护问题，提出了一种基于强化学习的方法RePolicy来动态调用安全策略，以适应未见过的执行路径和变化的政策环境。

Abstract

Safeguarding language model agents requires assessing complete execution trajectories under context-dependent safety policies. Existing policy-aware safeguards mainly rely on prompting or supervised fine-tuning, limiting their ability to adapt to unseen trajectories and changing policy contexts. We propose RePolicy, an agent safeguard that learns safety-policy invocation through reinforcement learning. Given an agent trajectory and a dynamic policy library, RePolicy invokes the applicable policy and uses its content to produce a policy-grounded rationale and safety judgment. We construct PolicyTraj-20K to support supervised initialization, followed by GRPO with verifiable rewards and policy-context perturbation. Experiments across six agent safety benchmarks show that RePolicy achieves strong overall safety-detection performance and robust policy invocation under varying policy contexts.

Problem

• Safety-Policy
• Agent Safeguards
• Reinforcement Learning
• Execution Trajectories
• Policy Contexts

Innovation

• Reinforcement Learning
• Safety-Policy Invocation
• Dynamic Policy Library
• GRPO
• Policy-Context Perturbation