What problem does it solve? Choosing or assessing software architecture is error-prone: teams over-engineer simple apps with microservices and CQRS, or under-architect systems that genuinely need distribution. This Skill guides architecture decisions for new systems and audits existing codebases against a pattern catalog, always recommending the least architecture that meets the requirement. ## Core Features & Use Cases - Greenfield selection interview: Walks a decision tree covering structure, topology, data, presentation, and concurrency axes, composing one pick per axis into a labelled stack with the cost of every choice. - Refactoring review: Maps an existing codebase's actual architecture, finds mismatches using catalog review cues, and recommends targeted incremental moves rather than rewrites. - Decision recording: Persists outcomes as proposed decision files with a regenerated constitution, and supports ADR migration and drift observation at checkpoints. - Use Case: Ask whether a new payments service should be a monolith or microservices, and receive a composed stack table with per-axis rationale, costs, and a least-architecture check. ## Quick Start Ask the assistant to help you choose an architecture for a new service or review whether your current codebase's architecture still fits.
