review-architecture

Maintain an architecture review log with confidence tiers for components and layers.

1|Updated Dec 23, 2025
One-click install
npx skills add https://github.com/huangeddie/dotfiles --skill review-architecture-huangeddie
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-architecture
Source: https://github.com/huangeddie/dotfiles/tree/main/dot_agents/exact_packages/coding/exact_skills/review-architecture
Command: npx skills add https://github.com/huangeddie/dotfiles --skill review-architecture-huangeddie

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

It prevents architecture drift by giving you a consistent place to catalog and review data models, interfaces, and unit tests across abstraction layers using explicit confidence tiers.

Core Features & Use Cases

  • Architecture cataloging: Maintain a small docs/architecture directory that organizes reviewed components by layer and cross-cutting concern.
  • Confidence-based reviews: Record holistic layer and concern health with tiered ratings and last-reviewed dates, plus optional notes.
  • Component-level tracking: Track individual models, interfaces, and unit tests with per-component metadata (file, reviewed date, modified date) without inventing ratings.

Usage Example

When you start a new codebase, catalog layers (e.g., Domain Model, Application Services) and then periodically record layer-level reviews before and after major refactors.

Quick Start

Use the review-architecture skill to maintain and update docs/architecture/ by reviewing a specific layer or cataloging architectural components.

Frequently Asked Questions about review-architecture

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I track architecture health and prevent codebase drift?

Track architectural health by maintaining a stable docs/architecture layout to catalog layers, cross-cutting concerns, and interface contracts with explicit confidence ratings. This prevents architecture drift by providing a consistent place to review data models and unit tests across abstraction layers.

What is the best way to document interface contracts for a codebase review?

Document interface contracts by maintaining a small docs/architecture directory that organizes reviewed components by layer and cross-cutting concern. You record holistic layer and concern health using tiered confidence ratings, last-reviewed dates, and optional notes to structure the review.

How do I structure an architecture review log before a major refactoring?

Structure an architecture review log by cataloging layers like Domain Model or Application Services, then recording layer-level reviews before and after major refactors. This requires a stable docs/architecture layout with per-layer and per-concern frontmatter acting as the source of truth.

Can I track individual data models and unit tests without assigning them a confidence rating?

Yes, you can track individual data models, interfaces, and unit tests using per-component metadata like file path, reviewed date, and modified date. The process enforces strict separation between component reviews and layer reviews, cataloging components without inventing ratings.

Does maintaining an architecture catalog require a specific directory structure?

Yes, maintaining an architecture catalog requires a stable docs/architecture layout with per-layer and per-concern frontmatter as the source of truth. This strict directory structure is necessary to ensure rating scale consistency and proper separation between component and layer reviews.