| title | When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents |
|---|---|
| authors | Kaiyue Yang, Yuyan Bu, Jingwei Yi, Yuchi Wang, Biyu Zhou, Juntao Dai, Songlin Hu, Yaodong Yang |
| date | 2026-06-18 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2606.20023 |
| affiliation | Beijing Academy of Artificial Intelligence |
| upvotes | 3 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | As LLM agents increasingly select tools autonomously, their choices among tools with different privileges become safety-relevant. However, prior tool-selection studies focus on safety-agnostic metadata preferences, leaving privilege-sensitive choices underexplored. To address this gap, we study over-privileged tool selection, in which an agent selects or escalates to a higher-privilege tool despite a sufficient lower-privilege alternative. We introduce ToolPrivBench to evaluate whether agents choose higher-privilege tools despite sufficient lower-privilege alternatives, measuring both initial selection and escalation after transient tool failures. Across eight domains and five recurring risk patterns, we find that over-privileged tool selection is common among mainstream LLM agents and is further amplified by transient failures. We further find that general safety alignment does not reliably transfer to least-privilege tool choice, while prompt-level controls provide only limited mitigation under transient failures. We therefore introduce a privilege-aware post-training defense that teaches agents to prefer sufficient lower-privilege tools and escalate only when necessary. Our mitigation experiments show that this defense substantially reduces unnecessary high-privilege tool use while preserving general capabilities. |
cite
@misc{yang2026when,
title={When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents},
author={Kaiyue Yang and Yuyan Bu and Jingwei Yi and Yuchi Wang and Biyu Zhou and Juntao Dai and Songlin Hu and Yaodong Yang},
year={2026},
eprint={2606.20023},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2606.20023}
}