The field of multi-agent systems is rapidly evolving, with a growing need for standardized protocols to enable secure and efficient communication between decentralized AI agents. Recent research has focused on developing open and decentralized collaboration infrastructures, such as the Coral Protocol, to facilitate complex workflows and collective intelligence. Furthermore, there is a increasing awareness of the security challenges associated with multi-agent systems, including threats such as secret collusion, coordinated swarm attacks, and data poisoning. To address these challenges, new fields of research are emerging, including multi-agent security, which aims to secure networks of decentralized AI agents against threats that arise from their interactions. Noteworthy papers in this area include:

• The Coral Protocol, which introduces a standardized messaging format and secure team formation capabilities for dynamic agent collaboration.
• Open Challenges in Multi-Agent Security, which proposes a unified research agenda for securing agent systems and interaction environments.
• A survey of agent interoperability protocols, which examines four emerging protocols for integrating tools and coordinating tasks across heterogeneous systems.