arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.03213cs.CRcs.AI
收藏

当代理处理秘密:关于代理AI的保密计算调查

When Agents Handle Secrets: A Survey of Confidential Computing for Agentic AI

Department of Computing Imperial College London London, United Kingdom(计算系帝国理工学院伦敦英国)

Javad Forough, Marios Kogias, Hamed Haddadi

AI总结:

本文探讨了代理AI中保密计算的应用,分析了六个TEE平台的设计,提出了基于代理的威胁模型,并指出了六个开放挑战,旨在为生产级代理AI提供安全基础。

摘要:

代理AI系统,特别是基于大语言模型的代理,能够计划、调用工具、维护持久内存,并通过MCP和A2A等协议将任务委托给同僚代理,引入了与独立模型推理截然不同的威胁面。代理会积累敏感上下文,持有凭证,并在单个方无法完全控制的管道中运行,导致提示注入、上下文外泄、凭证盗窃和跨代理信息中毒。当前的防御措施完全基于软件栈,可以被足够特权的对手如被入侵的云运营商静默绕过。保密计算(CC)提供了一种基于硬件的替代方案:可信执行环境(TEEs)将代理代码和数据与特权系统软件隔离,而远程认证可实现分布式部署中的可验证信任。本文将设计空间分为四部分:(i)涵盖部署角色和性能权衡的六个TEE平台统一分类(Intel SGX、Intel TDX、AMD SEV-SNP、ARM TrustZone、ARM CCA和NVIDIA H100 CC);(ii)以代理为中心的威胁模型,覆盖感知、规划、内存、行动和协调层,并映射到九个安全目标;(iii)对基于CC的防御措施的比较调查,区分从单次调用推理转移的发现与需要新代理设计的发现;(iv)六个开放挑战,包括多跳代理链的复合认证和GPU-TEE在LLM规模下的性能。尽管几种硬件信任原始语似乎足够成熟以供定向部署,但尚无广泛确立的端到端框架将它们整合成一个连贯的安全基质用于生产代理AI。

英文摘要:

Agentic AI systems, specifically LLM-driven agents that plan, invoke tools, maintain persistent memory, and delegate tasks to peer agents via protocols such as MCP and A2A, introduce a threat surface that differs materially from standalone model inference. Agents accumulate sensitive context, hold credentials, and operate across pipelines no single party fully controls, enabling prompt injection, context exfiltration, credential theft, and inter-agent message poisoning. Current defenses operate entirely within the software stack and can be silently bypassed by a sufficiently privileged adversary such as a compromised cloud operator. Confidential computing (CC) offers a hardware-rooted alternative: Trusted Execution Environments (TEEs) isolate agent code and data from privileged system software, while remote attestation enables verifiable trust across distributed deployments. This survey synthesizes the design space in four parts: (i) a unified taxonomy of six TEE platforms (Intel SGX, Intel TDX, AMD SEV-SNP, ARM TrustZone, ARM CCA, and NVIDIA H100 CC) covering deployment roles and performance tradeoffs; (ii) an agent-centric threat model spanning perception, planning, memory, action, and coordination layers mapped to nine security goals; (iii) a comparative survey of CC-based defenses distinguishing findings that transfer from single-call inference versus what requires new agentic designs; and (iv) six open challenges including compound attestation for multi-hop agent chains and GPU-TEE performance at LLM scale. While several hardware trust primitives appear mature enough for targeted deployments, no broadly established end-to-end framework yet binds them into a coherent security substrate for production agentic AI.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑