Show HN: Claude-relais – A plan/build/judge loop mixing Claude with Cursor (github.com)

The newly announced "claude-relais" is an innovative orchestration skill that enhances the functionality of Claude by integrating it with Cursor and Guardrails for a more streamlined coding process. This tool adopts a distinct PLAN -> BUILD -> JUDGE framework, allowing Claude's advanced reasoning capabilities to focus on task design while Cursor handles execution efficiently. Users can navigate the orchestration with clear commands, leveraging Claude for high-level planning and Cursor for rapid implementation. The service is affordable, with estimates of around $40/month, depending on usage. This development is significant for the AI/ML community as it represents a noteworthy step towards optimizing the collaboration between large language models and agile execution tools. By separating the reasoning and execution roles, claude-relais improves throughput and safety during the coding process. It includes essential safety checks to prevent errors and supports a guided onboarding experience, making it accessible to users of varying expertise. With the integration of Git support and easy task management, it paves the way for enhanced productivity in code orchestration.