# Architecture Review

> Analyze module architecture using polyscan - class coupling (CBO), class cohesion (LCOM4), instability metrics, and circular dependency detection. Use when user asks about architecture, module structure, coupling, or circular dependencies.

- Skill: `ludo-technologies/architecture-review` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ludo-technologies/architecture-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ludo-technologies/architecture-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools, AI & ML
- Author: ludo-technologies (https://skillmd.com/u/ludo-technologies)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/ludo-technologies/architecture-review

---


# Architecture Review with polyscan

Run the polyscan CLI to understand module structure and coupling. Dependency analysis exists for Go (package graph) and JavaScript/TypeScript (module graph); coupling (CBO) for Go, Rust and JavaScript/TypeScript; class cohesion (LCOM4) for Go and Rust; C++ gets complexity and clone detection only. No install needed: `npx polyscan@latest analyze <path>`.

## Commands

| User Request | Command |
|-------------|---------|
| "Check class coupling" | `npx polyscan@latest analyze --format text --select cbo <path>` |
| "Which Go or Rust types do too many things?" | `npx polyscan@latest analyze --format text --select lcom <path>` |
| "Find circular dependencies" | `npx polyscan@latest analyze --format text --select deps <path>` |
| "Which modules are risky to change?" | `npx polyscan@latest analyze --format json --select deps,cbo <path>` |

Text output shows per-class CBO and cycle membership, but only **aggregate** coupling stats for modules. Per-module Martin metrics (Ca/Ce, instability, abstractness, distance, risk level) and the explicit Zone of Pain / main-sequence module lists live in the `deps` section of the `--format json` output, which goes to stdout with the health-score summary on stderr.

## Interpreting Coupling Results

- High CBO classes depend on many others; changes ripple widely. Suggest interface extraction or dependency inversion. For Go and Rust the count covers types declared in the analyzed tree only, so a type coupled only to the standard library reads as 0 and is not listed.
- A high LCOM4 type has methods that fall into several groups sharing no field or call; each group is a candidate for its own type.
- Martin metrics per module (from the JSON output above): instability I = Ce / (Ca + Ce) and distance from the main sequence. Modules in the Zone of Pain (stable but concrete) are risky to change; name them explicitly.
- Dependency cycles are the highest-priority architectural issue; name the modules in each cycle and the weakest edge to break.

Always tie findings back to concrete modules and suggest a specific structural change.