Show HN: Covenant – A governance framework for multi-agent AI systems (github.com)

Covenant has unveiled a robust governance framework designed for managing multi-agent AI systems, addressing the critical coordination challenges that arise when multiple AI agents collaborate. This framework provides structured lifecycle management and quality controls to ensure seamless communication, rule adherence, and efficient recovery from failures among agents. The system operates within the Claude Code or OpenAI Codex CLI, requiring no external dependencies, making it easily accessible for developers looking to build complex AI-driven applications. Significantly, Covenant simplifies the development process by allowing users to describe their project needs, which the framework then translates into a governed project with executable agent code and structured memory management. With features such as exit reports and a memory system to retain knowledge across sessions, developers can expect increased efficiency and consistency in project execution. The community edition is free, offering a wide range of tools, from application scaffolding to agent-native programming languages, making it an essential resource for AI/ML professionals eager to leverage the power of multi-agent systems without the typical overhead of configuration and setup.