I have built a supervision and Orchestration layer for Claude Code and other CLI (cli.agentproto.sh)

A new supervision and orchestration layer for Claude Code and other command-line interfaces (CLIs) has been announced, offering developers a comprehensive tool to manage and monitor their AI applications more effectively. This innovative system, named AgentProto, allows users to run agents, manage sessions, and orchestrate multiple operations from a single binary. Its ability to serve live sessions through a unified dashboard provides visibility into every interaction and event, making it easier for developers to troubleshoot and optimize their workflows. The significance of AgentProto lies in its adaptability and integration capabilities, supporting various adapters and enabling seamless communication with local and remote daemons via a WebSocket tunnel. Users can install specific adapters like Claude Code or Hermes, run agents either for single turns or persistent sessions, and orchestrate complex infrastructures through composed manifests. These features enhance collaboration within the AI/ML community by streamlining integrations with existing technologies and providing detailed documentation to guide developers in implementing this orchestration layer into their projects.