SoK: Trust-Authorization Mismatch in LLM Agent Interactions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shi, Guanquan, Du, Haohua, Wang, Zhiqiang, Liang, Xiaoyu, Liu, Weiwenpei, Bian, Song, Guan, Zhenyu
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914315262492672
author Shi, Guanquan
Du, Haohua
Wang, Zhiqiang
Liang, Xiaoyu
Liu, Weiwenpei
Bian, Song
Guan, Zhenyu
author_facet Shi, Guanquan
Du, Haohua
Wang, Zhiqiang
Liang, Xiaoyu
Liu, Weiwenpei
Bian, Song
Guan, Zhenyu
contents Large Language Models (LLMs) are evolving into autonomous agents capable of executing complex workflows via standardized protocols (e.g., MCP). However, this paradigm shifts control from deterministic code to probabilistic inference, creating a fundamental Trust-Authorization Mismatch: static permissions are structurally decoupled from the agent's fluctuating runtime trustworthiness. In this Systematization of Knowledge (SoK), we survey more than 200 representative papers to categorize the emerging landscape of agent security. We propose the Belief-Intention-Permission (B-I-P) framework as a unifying formal lens. By decomposing agent execution into three distinct stages-Belief Formation, Intent Generation, and Permission Grant-we demonstrate that diverse threats, from prompt injection to tool poisoning, share a common root cause: the desynchronization between dynamic trust states and static authorization boundaries. Using the B-I-P lens, we systematically map existing attacks and defenses and identify critical gaps where current mechanisms fail to bridge this mismatch. Finally, we outline a research agenda for shifting from static Role-Based Access Control (RBAC) to dynamic, risk-adaptive authorization.
format Preprint
id arxiv_https___arxiv_org_abs_2512_06914
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SoK: Trust-Authorization Mismatch in LLM Agent Interactions
Shi, Guanquan
Du, Haohua
Wang, Zhiqiang
Liang, Xiaoyu
Liu, Weiwenpei
Bian, Song
Guan, Zhenyu
Cryptography and Security
Artificial Intelligence
Large Language Models (LLMs) are evolving into autonomous agents capable of executing complex workflows via standardized protocols (e.g., MCP). However, this paradigm shifts control from deterministic code to probabilistic inference, creating a fundamental Trust-Authorization Mismatch: static permissions are structurally decoupled from the agent's fluctuating runtime trustworthiness. In this Systematization of Knowledge (SoK), we survey more than 200 representative papers to categorize the emerging landscape of agent security. We propose the Belief-Intention-Permission (B-I-P) framework as a unifying formal lens. By decomposing agent execution into three distinct stages-Belief Formation, Intent Generation, and Permission Grant-we demonstrate that diverse threats, from prompt injection to tool poisoning, share a common root cause: the desynchronization between dynamic trust states and static authorization boundaries. Using the B-I-P lens, we systematically map existing attacks and defenses and identify critical gaps where current mechanisms fail to bridge this mismatch. Finally, we outline a research agenda for shifting from static Role-Based Access Control (RBAC) to dynamic, risk-adaptive authorization.
title SoK: Trust-Authorization Mismatch in LLM Agent Interactions
topic Cryptography and Security
Artificial Intelligence
url https://arxiv.org/abs/2512.06914