Show HN: Cross-agent messaging and shared memory over the local filesystem (oacp.dev)

A new protocol called OACP (Open Agent Coordination Protocol) has been introduced to enhance multi-agent AI workflows by enabling cross-agent messaging and persistent shared memory using a simple file-based system. This innovative approach allows agents, such as Claude, Codex, and Gemini, to communicate asynchronously through typed messages and manage shared data across various projects and machines, all without the need for a central server or complex infrastructure. The protocol comprises specific message types—like task requests and review feedback—and organizes interactions in structured review loops to meet quality standards before finalizing changes. This development holds significant potential for the AI and machine learning community, as it simplifies coordination among different AI agents and facilitates seamless collaboration across diverse platforms. By leveraging plain YAML and markdown files for communication, OACP promotes an open and self-hostable model that minimizes dependencies. Its runtime-agnostic nature ensures compatibility with any agent capable of reading and writing files, making it a versatile tool for developers aiming to streamline multi-agent interactions and maintain project consistency across varying environments.