This skill provides a comprehensive framework for refactoring complex 'god objects' into manageable, isolated domains within the Claude Flow ecosystem. By leveraging the microkernel pattern and clean architecture layers—Presentation, Application, Domain, and Infrastructure—it enables developers to build highly testable, decoupled, and scalable AI agent systems. It is specifically designed for enterprise-grade projects that require strict separation of concerns and robust domain logic to manage autonomous workflows effectively.
主な機能
01Supports a microkernel-based plugin system for modular extensions
0212,331 GitHub stars
03Implements Clean Architecture layers to ensure zero domain dependencies
04Facilitates event-driven communication via standardized domain events
05Provides standardized testing patterns for pure domain logic validation
06Decomposes monolithic orchestrators into isolated bounded contexts
ユースケース
01Refactoring large AI agent orchestrators into maintainable domain modules
02Designing scalable multi-agent systems with clear boundaries and shared value objects
03Implementing a plugin-based architecture to extend agent capabilities without core modification