Architecture Reviewer FAQs

Question 1

Which design patterns does this skill enforce?

Accepted Answer

It primarily enforces the Template Method pattern for experiment Runners (via the AbstractRunner class) and the Strategy pattern for interchangeable training algorithms (via the AbstractTrainer class).

Question 2

What does the Architecture Reviewer skill do?

Accepted Answer

The Architecture Reviewer skill validates code against the PRISM architectural framework. It specifically ensures that new or refactored Runner and Trainer classes adhere to defined Template Method and Strategy patterns to maintain a clean, modular codebase.

Question 3

Does it support type hinting and naming conventions?

Accepted Answer

Yes, the Architecture Reviewer verifies proper Python type hints, validates specific return types like ExperimentResult, and enforces consistent naming standards such as PascalCase for classes and snake_case for methods.

Question 4

When should I use this Claude Code skill?

Accepted Answer

This skill is best used during refactoring phases, when implementing new experiment workflows, or when reviewing pull requests. It ensures that any structural changes align with the project's long-term design goals and architectural standards.

Question 5

How does it help prevent code anti-patterns?

Accepted Answer

The skill identifies common architectural smells, such as overriding protected template methods, mixing concerns (like putting visualization logic inside a Trainer), and hardcoding file paths instead of using configuration objects.

Architecture Reviewer

Architecture Reviewer

Key Features

Use Cases

Key Features

Use Cases