This work constructs a threat model enumerating 14 attack vectors, finding that 67% of agents are vulnerable to at least one scope violation even with system-prompt-level guardrails, and indirect injection via tool outputs succeeds in 43% of attempts.
Abstract
Existing prompt injection research focuses on single-model chatbot scenarios, where an attacker manipulates one LLM through crafted input. Multi-agent systems amplify this threat through three mechanisms absent from single-model settings: inter-agent message passing creates injection channels invisible to perimeter defenses, shared tool access enables privilege escalation across agent boundaries, and trust propagation allows a compromised agent to influence upstream orchestrators. We construct a threat model enumerating 14 attack vectors across four categories: direct injection via user input (3 vectors), indirect injection via tool outputs (4 vectors), inter-agent injection via message passing (4 vectors), and cascading injection through orchestrator manipulation (3 vectors). Testing all 14 vectors against a 6-agent production-representative system, we find that 67% of agents are vulnerable to at least one scope violation even with system-prompt-level guardrails, and indirect injection via tool outputs succeeds in 43% of attempts. Four architectural defenses reduce overall injection success from 31.2% to 4.2%: message signing with provenance tracking (inter-agent injection down 91%), input/output sanitization at agent boundaries (indirect injection down 78%), privilege-scoped tool access per agent role (privilege escalation eliminated entirely), and anomaly detection on inter-agent communication patterns (84% of cascading attempts caught).
This work constructs a concrete prompt-injection attack that succeeds despite all constituent agents individually operating CaMeL, and introduces multi-CaMeL, an agent-to-agent communication protocol that preserves provenance across agent boundaries by separating trusted natural-language instructions from untrusted dat...
James Peters-Gill, Avi Semler, Henning Bartsch et al.· 3 citations
The findings show that defense evaluation should report attack outcomes, goal feasibility, legitimate-task utility, action availability, operational measures, and run-to-run variation.
Adil Khan, Khaled AlKhanbashi, Azza Mohamed· Computers· 2 citations
Two universal tool-based defenses are introduced: Attacker Tool Filtering, which uses anomaly detection to identify and remove suspicious tools, and Normal Tool Recalling, a white-box method that restores the agent's original toolset prior to planning.
A threat model for agentic IaC systems in the form of the structure SIaC = (Agents, Environment, Configurations, Outcomes), which makes it possible to identify the critical points of influence: agent input data, configuration state, agent intents and goals, inter-agent communication channels, the execution process, aut...
Mykhailo Savonik, I. Parkhomenko· Cybersecurity: Education, Sc...· 0 citations
ProtoArena is introduced, a black-box, multi-component, dynamic benchmark with 80 cross-component vulnerability-exploitation environments spanning 11 protocol and application domains and 7 attack categories and success with limited black-box information remains low and varies across domains and attack categories.
Fa-Sheng Miao, Zhao-Yu Hu, Xiang Li et al.· 0 citations
A systematic review and structured descriptive synthesis of research on defenses against prompt-based attacks in language model and agent systems reveals trade-offs between security effectiveness, performance, and system complexity as well as major gaps in benchmarks, indirect attack coverage, and multi-agent evaluatio...
Sana Mourad, E. E. Abdallah, Mohammad Ababneh· Electronics· 0 citations