arXiv · 2611.11539
PRISM: Causal Counterfactual Minimization of Agent Tool Privileges via Information Bottleneck Role Induction
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, $τ$-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%.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Yang Luo, Qingni Shen, Tairui Wang, Jiajun Li, Zhonghai Wu. 2026-11-08. PRISM: Causal Counterfactual Minimization of Agent Tool Privileges via Information Bottleneck Role Induction. https://searcharxiv.com/abs/2611.11539
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.