Can it detect circular dependencies?

Yes, it specifically analyzes the dependency graph to identify circular imports and inappropriate coupling between high-level and low-level modules.

What is the Architecture Review skill?

It is a specialized capability for Claude Code that focuses on high-level design, structural integrity, and the architectural fit of code changes within a project.

How does this differ from a standard code review?

While standard reviews focus on logic and bugs, this skill acts as a tech lead, evaluating abstraction levels, module boundaries, and long-term maintainability impact.

Does it support custom project conventions?

Yes, it performs deep context research by reading files like CLAUDE.md and project configs to align its review with your specific architectural rules.

Architecture & Design Review

Name: Architecture & Design Review
Author: liminal-ai

byliminal-ai

0•

コラボレーションツール

Evaluates code changes for architectural fit, design patterns, and structural integrity to ensure long-term maintainability.

This skill acts as a virtual tech lead, examining code changes not just for functionality, but for their impact on the broader codebase structure. It helps developers maintain consistent abstraction levels, manage complex dependencies, and ensure that new features align with established design patterns. By conducting deep context research and analyzing module boundaries, it identifies architectural risks such as tight coupling, leaky abstractions, and circular dependencies before they become technical debt.

主な機能

01Scorecard-based assessment of architectural conformance and maintainability

02Analysis of coupling, dependency direction, and circular references

03Detection of leaky abstractions and layer violations across the codebase

04Deep context research of existing module structures and directory conventions

05Evaluation of API surface clarity and public contract consistency

060 GitHub stars

ユースケース

01Performing a high-level 'tech lead' pass on complex pull requests

02Validating that a new feature fits existing project design patterns

03Refactoring legacy modules to improve separation of concerns

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add liminal-ai/code-steward architecture-review

For use in Claude.ai and ChatGPT

主な機能

01Scorecard-based assessment of architectural conformance and maintainability

02Analysis of coupling, dependency direction, and circular references

03Detection of leaky abstractions and layer violations across the codebase

04Deep context research of existing module structures and directory conventions

05Evaluation of API surface clarity and public contract consistency

060 GitHub stars

ユースケース

01Performing a high-level 'tech lead' pass on complex pull requests

02Validating that a new feature fits existing project design patterns

03Refactoring legacy modules to improve separation of concerns

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add liminal-ai/code-steward architecture-review

For use in Claude.ai and ChatGPT