layered-rails

Extract an operation model callback to an event-driven subscriber using ActiveSupport::Notifications.

Extract UI-specific validations and callbacks from a model/controller into a form object.

Extract inline authorization logic from controllers into a '<Model>Policy' object with standardized constructor and predicate methods.

Extract presentation logic from a model into a properly structured presenter object.

Extract a complex query method or scope from a model into a query object following The Extended Rails Way conventions.

Use this agent when a developer wants to progressively introduce abstraction layers into an existing Rails codebase. Helps with: sequencing a multi-step adoption plan, deciding which layers to introduce first using churn×complexity analysis, recommending extraction order (callbacks → service objects → query objects → form objects → presenters → policy objects → notifications → view components → configuration objects), respecting the gradual-refactoring principle (localize first, then improve), and avoiding premature abstraction. Applies The Extended Rails Way philosophy from Layered Design for Ruby on Rails Applications.

Use this agent when you want a code review of Rails code against The Extended Rails Way's anti-patterns and layer rules. Checks for: God object models, fat controllers, callback accumulation, concerns as code splitters, presentation logic in models, leaking decorators, direct third-party injection, ENV hell, Rails.env checks in application code, instance variable leakage in partials, authorization logic in wrong layers, anemic models, N+1 authorization, raw SQL stitching, over-scoping, notification delivery in wrong layers, puts-based logging, vendor-specific exception APIs, and more. Provides specific remediation guidance pointing to the relevant pattern references.

Guides Rails developers applying The Extended Rails Way: gradually extracting abstraction layers (service objects, query objects, form objects, presenters, policy objects, and more) to keep growing Rails codebases maintainable. Use when reviewing Rails code for layered architecture violations, identifying God object models, extracting patterns from fat controllers or fat models, or planning progressive refactoring of an existing Rails application.