PSG-Agent：面向基于 LLM 的智能体的人格感知安全护栏
PSG-Agent: Personality-Aware Safety Guardrail for LLM-based Agents
AI总结：
针对现有智能体护栏忽视用户差异与多轮风险累积的问题，提出 PSG-Agent，结合稳定人格特质、实时状态及全流程专项监控实现个性化动态安全防护。
摘要：
有效的护栏对于在关键应用中安全部署基于 LLM 的智能体至关重要。尽管近期取得了进展，现有护栏仍存在两个根本局限：（i）它们对所有用户应用统一的护栏策略，忽视了同一智能体行为可能对某些用户造成伤害，而对其他用户却是安全的；（ii）它们孤立地检查每条响应，遗漏了风险在多次交互中如何演变和累积。为解决这些问题，我们提出 PSG-Agent，这是一个面向基于 LLM 的智能体的个性化动态系统。首先，PSG-Agent 通过挖掘交互历史中的稳定特质并从当前查询中捕捉实时状态来创建个性化护栏，生成针对特定用户的风险阈值和保护策略。其次，PSG-Agent 通过专门的防护组件在整个智能体流水线中实施持续监控，包括 Plan Monitor、Tool Firewall、Response Guard 和 Memory Guardian，它们跟踪跨轮次风险累积并给出可验证的判定。最后，我们在医疗、金融和日常生活自动化等多种场景下，结合多样化用户画像对 PSG-Agent 进行验证。它显著优于包括 LlamaGuard3 和 AGrail 在内的现有智能体护栏，为基于 LLM 的智能体的个性化安全提供了一条可执行、可审计的路径。
英文摘要：
Effective guardrails are essential for safely deploying LLM-based agents in critical applications. Despite recent advances, existing guardrails suffer from two fundamental limitations: (i) they apply uniform guardrail policies to all users, ignoring that the same agent behavior can harm some users while being safe for others; (ii) they check each response in isolation, missing how risks evolve and accumulate across multiple interactions. To solve these issues, we propose PSG-Agent, a personalized and dynamic system for LLM-based agents. First, PSG-Agent creates personalized guardrails by mining the interaction history for stable traits and capturing real-time states from current queries, generating user-specific risk thresholds and protection strategies. Second, PSG-Agent implements continuous monitoring across the agent pipeline with specialized guards, including Plan Monitor, Tool Firewall, Response Guard, Memory Guardian, that track cross-turn risk accumulation and issue verifiable verdicts. Finally, we validate PSG-Agent in multiple scenarios including healthcare, finance, and daily life automation scenarios with diverse user profiles. It significantly outperform existing agent guardrails including LlamaGuard3 and AGrail, providing an executable and auditable path toward personalized safety for LLM-based agents.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
