Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.23614 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908563040894976 |
|---|---|
| author | Wu, Yaozu Guo, Jizhou Li, Dongyuan Zou, Henry Peng Huang, Wei-Chieh Chen, Yankai Wang, Zhen Zhang, Weizhi Li, Yangning Zhang, Meng Jiang, Renhe Yu, Philip S. |
| author_facet | Wu, Yaozu Guo, Jizhou Li, Dongyuan Zou, Henry Peng Huang, Wei-Chieh Chen, Yankai Wang, Zhen Zhang, Weizhi Li, Yangning Zhang, Meng Jiang, Renhe Yu, Philip S. |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_23614 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | PSG-Agent: Personality-Aware Safety Guardrail for LLM-based Agents Wu, Yaozu Guo, Jizhou Li, Dongyuan Zou, Henry Peng Huang, Wei-Chieh Chen, Yankai Wang, Zhen Zhang, Weizhi Li, Yangning Zhang, Meng Jiang, Renhe Yu, Philip S. Artificial Intelligence 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. |
| title | PSG-Agent: Personality-Aware Safety Guardrail for LLM-based Agents |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2509.23614 |