arch-assess

Start by reading the code to build a module dependency graph, then reverse-engineer as-is C4 Context, Container, and Component diagrams. Next, locate coupling hotspots and architecture smells, and finally score quality attributes like scalability and maintainability on a 1-5 scale to produce a health report.

Use static analysis tools such as madge or dependency-cruiser for TypeScript, pydeps for Python, and go mod graph for Go to build dependency graphs. Then rank modules by fan-in; the top fan-in modules combined with high change frequency are the highest-risk coupling hotspots.

Key structural smells include distributed monoliths with strong synchronous calls, shared databases across services, god services with too many responsibilities, and circular dependencies. Data smells include shared mutable state and missing transaction boundaries, while deployment smells include single-instance deployment and hardcoded configuration.

Use assessment when working with an existing brownfield system to diagnose its current state before planning changes. Use architecture design for greenfield systems built from scratch, architecture review for evaluating proposed solutions, and evolution planning only after the assessment diagnosis is complete.

No, the assessment deliberately separates diagnosis from prescription. It outputs a health report with scored quality attributes, coupling hotspots, and a prioritized smell list, which then feeds into a separate architecture evolution workflow that designs the improvement roadmap.