Agent Skillsandonimichael/arxitect › clean-architecture-review

clean-architecture-review

GitHub

用于审查代码是否符合整洁架构规范,评估组件内聚、耦合及可维护性等质量属性。

skills/clean-architecture-review/SKILL.md andonimichael/arxitect

Trigger Scenarios

审查新代码或修改代码的架构合规性 评估系统架构设计的合理性

Install

npx skills add andonimichael/arxitect --skill clean-architecture-review -g -y
More Options

Use without installing

npx skills use andonimichael/arxitect@clean-architecture-review

指定 Agent (Claude Code)

npx skills add andonimichael/arxitect --skill clean-architecture-review -a claude-code -g -y

安装 repo 全部 skill

npx skills add andonimichael/arxitect --all -g -y

预览 repo 内 skill

npx skills add andonimichael/arxitect --list

SKILL.md

Frontmatter
{
    "name": "clean-architecture-review",
    "description": "Reviews code for clean architecture compliance including component cohesion principles (REP, CRP, CCP), component coupling principles (ADP, SDP, SAP), and quality attributes (maintainability, extensibility, testability). Use when evaluating architectural soundness of new or modified code."
}

Clean Architecture Review

You are performing a clean architecture review. Evaluate the code against three dimensions: component cohesion, component coupling, and quality attributes.

Review Process

  1. Identify the scope. Determine the files you are meant to review. Either the files specified, recently changed files, or the entire codebase.

  2. Map the component structure. Identify the logical components (packages, modules, directories) and their dependency relationships. Trace import/require statements to build a dependency graph.

  3. Evaluate component cohesion. For each component, assess compliance with the three cohesion principles. See component-cohesion.md for detailed evaluation criteria.

  4. Evaluate component coupling. Analyze dependencies between components for direction, stability, and abstraction. See component-coupling.md for detailed evaluation criteria.

  5. Assess quality attributes. Evaluate testability, extensibility, and maintainability of the design. See quality-attributes.md for specific indicators.

  6. Produce structured output. Follow the review output format defined in skills/architect/review-output-format.md. Every finding must include a severity, the principle violated, affected files, and a specific recommendation.

Severity Guidelines

  • CRITICAL: Dependency cycles exist, stable components depend on unstable ones, layer boundaries are violated, or the architecture prevents independent testing of core business logic.
  • WARNING: Component boundaries are unclear, a component mixes concerns that change at different rates, or dependencies could be better organized.
  • SUGGESTION: Minor reorganization that would improve clarity or make the architecture more self-documenting.

Pragmatism

Architecture must serve the system's actual scale. A single-module utility does not need hexagonal architecture. A two-file script does not need dependency inversion layers. Evaluate architectural decisions against the complexity they manage, not against an ideal textbook diagram.

The goal is an architecture that makes the system easy to understand, change, test, and deploy independently -- not one that wins an architecture diagram contest.

Version History

  • 473c486 Current 2026-07-25 04:15

Same Skill Collection

skills/api-design-review/SKILL.md
skills/architect/SKILL.md
skills/architecture-review/SKILL.md
skills/oo-design-review/SKILL.md
skills/using-arxitect/SKILL.md

Metadata

Files
0
Version
473c486
Hash
fecf6df2
Indexed
2026-07-25 04:15

inicio - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-08-22 04:30
浙ICP备14020137号-1 $mapa de visitantes$