Automates a three-stage architectural design process by chaining specialized AI subagents for creation, refinement, and critical review.
The Architect-Refine-Critique skill orchestrates a rigorous design review cycle by sequentially invoking three specialized subagents from within the Claude Code environment. First, the Architect generates an initial design for a specified target; second, the Refiner improves the design's efficiency and clarity; and third, the Critique subagent challenges the proposal to identify potential flaws or edge cases. This structured workflow ensures that technical designs are thoroughly vetted from multiple perspectives before implementation begins, saving developers time on manual context-switching.
주요 기능
01Rigorous critical analysis of technical proposals
02Sequential three-phase agent chaining
03Automated initial architectural design drafting
04Seamless hand-off for final design reviews
05Iterative design refinement and optimization
06121 GitHub stars
사용 사례
01Validating high-stakes technical decisions through adversarial AI critique
02Refactoring legacy codebases with multi-layered review requirements
03Designing complex software architectures or system components