Preprint
Reinforcement Learning

AI Agent Communication Protocols

Dezhang Kong, Shi Lin, Zhenhua Xu, Zhebo Wang, Minghao Li, Yufeng Li, Yilun Zhang, Hujin Peng, Zeyang Sha, Yuyuan Li, Changting Lin, Xun Wang, Xuan Liu, Ningyu Zhang, Chao-Jun Chen, Muhammad Khurram Khan, Meng Han
June 24, 2025arXiv.org71 citations
arXivPDFDOI

71

Citations

6

Influential Citations

arXiv.org

Venue

2025

Year

Abstract

In recent years, Large-Language-Model-driven AI agents have exhibited unprecedented intelligence and adaptability. Nowadays, agents are undergoing a new round of evolution. They no longer act as an isolated island like LLMs. Instead, they start to communicate with diverse external entities, such as other agents and tools, to perform complex tasks. Under this trend, agent communication is regarded as a foundational pillar of the next communication era, and many organizations have intensively begun to design related communication protocols (e.g., Anthropic's MCP and Google's A2A) within the past year. However, this new field exposes significant security hazards, which can cause severe damage to real-world scenarios. To help researchers quickly figure out this promising topic and benefit the future agent communication development, this paper presents a comprehensive survey of agent communication security. More precisely, we present the first clear definition of agent communication. Besides, we propose a framework that categorizes agent communication into three classes and uses a three-layered communication architecture to illustrate how each class works. Next, for each communication class, we dissect related communication protocols and analyze the security risks, illustrating which communication layer the risks arise from. Then, we provide an outlook on the possible defense countermeasures for each risk. In addition, we conduct experiments using MCP and A2A to help readers better understand the novel vulnerabilities brought by agent communication. Finally, we discuss open issues and future directions in this promising research field. We also publish a repository that maintains a list of related papers on https://github.com/theshi-1128/awesome-agent-communication-security.

Analysis

Why This Paper Matters

As LLM-driven AI agents evolve from isolated models to interconnected entities communicating with tools and other agents, the need for secure communication protocols becomes critical. This paper addresses a timely and underexplored area: the security risks inherent in agent communication. With major organizations like Anthropic and Google releasing protocols such as MCP and A2A, understanding vulnerabilities is essential to prevent real-world damage. The paper's comprehensive survey and framework provide a structured way to think about these risks, making it a valuable resource for researchers and practitioners entering this field.

Technical Contributions

  • Definition and Taxonomy: The paper provides the first clear definition of agent communication and proposes a three-class categorization (e.g., agent-to-agent, agent-to-tool, agent-to-human) along with a three-layered architecture (transport, protocol, application).
  • Security Analysis: For each communication class, the paper dissects related protocols and identifies specific security risks, mapping them to the communication layer from which they originate.
  • Experimental Validation: The authors conduct experiments using MCP and A2A to demonstrate novel vulnerabilities, offering concrete examples of how attacks can manifest.
  • Defense Outlook: The paper discusses potential countermeasures for each identified risk, providing a roadmap for future secure protocol design.

Results

The paper does not report quantitative metrics such as attack success rates or performance benchmarks. Instead, its results are qualitative: a structured taxonomy of security risks, identification of vulnerabilities in MCP and A2A, and a list of defense strategies. The experimental section illustrates specific attack scenarios but lacks numerical evaluation.

Significance

This survey fills a critical gap in the AI safety literature by focusing on the communication layer of agent systems. As multi-agent systems become more prevalent, the security of inter-agent communication will be paramount. The paper's framework can inform the design of future protocols and inspire further research into defenses. The public repository of related papers also serves as a valuable community resource. However, the lack of empirical benchmarks limits its immediate applicability for practitioners seeking quantitative guidance.

ai-agentscommunication-protocolssecuritylarge-language-modelssurveyreinforcement-learning

Related Papers