cs.CRNov 8, 2026

PRISM: Causal Counterfactual Minimization of Agent Tool Privileges via Information Bottleneck Role Induction

Authors: Yang Luo, Qingni Shen, Tairui Wang, Jiajun Li, Zhonghai Wu

Organizations: Peking University

Abstract

This paper shifts the object of role engineering from human users to large language model (LLM) agents, and proposes PRISM to address the over-authorization that arises when an agent obtains full tool privileges through the Model Context Protocol (MCP). PRISM formulates whether a tool is necessary for a task as an intervenable causal proposition, and applies structured interventions in a sandbox to learn a counterfactual necessity model. It then compresses trajectories into discrete latent variables through an information bottleneck, and determines the number of roles from the knee of the rate-distortion curve. Finally, it formalizes pruning as a bilevel Stackelberg game, in which a generative flow network samples diverse near-optimal solutions. On three public benchmarks, AgentDojo, ττ-bench and ToolEmu, PRISM compresses the number of authorized tools per role by 78.3%, attains a task success rate of 78.2%, reaches a counterfactual necessity accuracy of 88.7%, and lowers the blast radius to 15.9%.

Figures & tables

Explore similar work

CardsList
  1. When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents

    Jun 18, 2026Kaiyue Yang, Yuyan Bu, Jingwei Yi +5LLM Agent SecurityAI Agent Security Benchmarks

  2. OverAct: Measuring and Mitigating Proactive Over-Authorization in LLM Tool-Calling Agents

    Oct 1, 2026Taolin Zhang, Jiuheng Wan, Hanyu Wang +2Privacy Leakage in Language ModelsAI Agent Security Benchmarks