Secure and Governed Multi-Agent AI Architectures Using Identity-Aware Authorization and Fine-Grained Access Control
DOI:
https://doi.org/10.15662/IJEETR.2026.0806005Keywords:
Multi-Agent Artificial Intelligence, Identity-Aware Authorization, Fine-Grained Access Control, Agentic AI Security, Zero Trust Architecture, Attribute-Based Access Control (ABAC), Role-Based Access Control (RBAC), Policy-Based Access Control, Agent-to-Agent Communication, Least-Privilege Access, Enterprise AI Governance, Workload Identity, Secure Tool Invocation, Delegated Authorization, Autonomous AI SystemsAbstract
The growing adoption of multi-agent artificial intelligence (AI) systems within enterprise environments has introduced new challenges in securing autonomous interactions, controlling access to sensitive resources, and maintaining governance across distributed computational workflows. Unlike conventional AI applications, multi-agent architectures involve multiple autonomous entities that independently interpret tasks, exchange contextual information, invoke external tools, and coordinate decisions across heterogeneous enterprise systems. These capabilities introduce complex security risks associated with unauthorized resource access, excessive agent privileges, identity impersonation, unintended information disclosure, and insufficient accountability for machine-generated actions. Traditional role-based authorization models and perimeter-oriented security mechanisms are often inadequate for managing the dynamic relationships and contextual dependencies inherent in autonomous agent ecosystems.
This article presents a generalized architectural framework for designing secure and governed multi-agent AI systems through identity-aware authorization, fine-grained access control, and policy-driven governance mechanisms. The proposed approach integrates workload identity management, attribute-based access control, policy-based authorization, least-privilege enforcement, and continuous trust evaluation to establish verifiable security boundaries between agents, tools, data resources, and enterprise services. Particular emphasis is placed on securing agent-to-agent communication, enforcing context-sensitive permissions during tool execution, managing delegated authority, and establishing auditable decision pathways across distributed workflows. The architecture further incorporates centralized policy decision capabilities, distributed enforcement mechanisms, cryptographic identity verification, and continuous monitoring to support scalable governance without unnecessarily restricting agent autonomy. The study examines architectural components, authorization workflows, operational security considerations, and evaluation criteria for implementing trustworthy multi-agent platforms. By combining identity-centric security principles with adaptive authorization and governance controls, the proposed framework provides a foundation for developing enterprise AI ecosystems that balance autonomous execution, operational scalability, regulatory accountability, and protection of organizational information assets.
References
[1] S. Rose, O. Borchert, S. Mitchell, and S. Connelly, "Zero Trust Architecture," National Institute of Standards and Technology, NIST Special Publication 800-207, Aug. 2020.
[2] V. C. Hu, D. Ferraiolo, D. R. Kuhn, A. Schnitzer, K. Sandlin, R. Miller, and K. Scarfone, "Guide to Attribute Based Access Control (ABAC) Definition and Considerations," National Institute of Standards and Technology, NIST Special Publication 800-162, 2014, updated Aug. 2019.
[3] Joint Task Force, "Security and Privacy Controls for Information Systems and Organizations," National Institute of Standards and Technology, NIST Special Publication 800-53, Revision 5, Sept. 2020.
[4] M. Jones, A. Nadalin, B. Campbell, J. Bradley, and C. Mortimore, "OAuth 2.0 Token Exchange," Internet Engineering Task Force, RFC 8693, Jan. 2020.
[5] E. Tabassi, "Artificial Intelligence Risk Management Framework (AI RMF 1.0)," National Institute of Standards and Technology, NIST AI 100-1, Jan. 2023.
[6] Z. Xi et al., "The Rise and Potential of Large Language Model Based Agents: A Survey," arXiv preprint arXiv:2309.07864, 2023.
[7] SPIFFE Project, "SPIFFE: Secure Production Identity Framework for Everyone," SPIFFE Specifications and Documentation.
[8] T. Lodderstedt, J. Bradley, A. Labunets, and D. Fett, "Best Current Practice for OAuth 2.0 Security," Internet Engineering Task Force, RFC 9700, BCP 240, Jan. 2025.
[9] OWASP Foundation, "OWASP Top 10 for Large Language Model Applications," 2025 Edition, 2025.





