Authorization Propagation in Multi-Agent AI Systems: Identity Governance as Infrastructure

Krti Tallam

arXiv:2605.05440 · 2026-05-08 공개 · arXiv · PDF

Abstract

The security discussion around agentic AI focuses heavily on prompt injection. This paper argues that multi-agent systems also create a distinct authorization problem: maintaining authorization invariants as non-human principals retrieve data, delegate tasks, and synthesize results across changing boundaries. We call this problem authorization propagation. It is not reducible to prompt injection and is not fully addressed by classical access-control models such as RBAC, ABAC, or ReBAC. The paper formalizes authorization propagation as a workflow-level property, identifies three sub-problems (transitive delegation, aggregation inference, and temporal validity), and derives seven structural requirements for authorization architectures in multi-agent AI systems. Recent work on invocation-bound capability tokens, task-scoped authorization envelopes, dependency-graph policy enforcement, and execution-count revocation demonstrates that the field is converging on the problem, but not yet on a complete architecture. The central claim is that identity governance must be treated as infrastructure: evaluated continuously, enforced at every interaction boundary, and designed into the system before orchestration logic is allowed to scale. Preliminary implementation evidence from a production enterprise AI platform shows that ordinary system behavior, not only adversarial action, already produces the failures this model predicts.

한국어 요약

📋 한 줄 요약

**[AI 보안/거버넌스]** 멀티에이전트 AI 시스템에서 prompt injection과는 구분되는 "권한 전파(authorization propagation)" 문제를 정의하고, 이를 워크플로우 수준 속성으로 형식화한 뒤 7가지 구조적 요건을 제안한다.

🎯 핵심 기여도

💡 핵심 아이디어

비인간 주체(에이전트)가 데이터 검색·작업 위임·결과 합성 단계마다 경계가 변하므로, 권한은 정적 정책이 아니라 워크플로우 전체에서 지속적으로 재평가·강제되는 인프라가 되어야 한다. 적대적 행동뿐 아니라 일상적 시스템 동작 자체도 같은 종류의 실패를 야기한다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 보안 논의를 prompt injection 위주에서 벗어나 멀티에이전트 시대의 핵심 인프라 문제로 재정의. **한계**: 포지션·요건 도출 중심으로 구체 구현 청사진이나 정량 평가 벤치마크는 향후 과제로 남는다.

🚀 실용적 활용