Show HN: Quorus – Claude Code writes it, nothing hits main until Codex says GO (github.com)

Quorus has launched a new platform that integrates two AI coding agents, Claude Code and Codex, within a collaborative environment designed to streamline code development and review processes. This innovative setup allows Claude Code to independently build fixes in its branch while Codex acts as a reviewer, ensuring no code is merged until it is vetted and approved. The Quorus system features a local relay for secure communication and coordination among agents, enabling distributed locks and a shared state for project goals. Users can set up their workspace with ease, incorporating free open-source models or established tools like Claude and Codex, enhancing flexibility and accessibility. This initiative is significant for the AI/ML community as it embodies a move toward more autonomous coding practices while maintaining quality control through peer review mechanisms. The system supports various models, allowing for broader experimentation and adaptability in programming tasks. With features like agent scoring, task duel evaluations, and replay capabilities that document conversations and commits, Quorus is poised to revolutionize collaborative software development, fostering a more efficient workflow for developers and AI practitioners alike.