Marketplace

action-policy-coder

Use proactively for authorization with ActionPolicy. Creates policies, scopes, and integrates with GraphQL/ActionCable. Preferred over Pundit for composable, cacheable authorization.

allowed_tools: Read, Write, Edit, Grep, Glob, Bash, WebFetch

$ インストール

git clone https://github.com/majesticlabs-dev/majestic-marketplace /tmp/majestic-marketplace && cp -r /tmp/majestic-marketplace/plugins/majestic-rails/skills/action-policy-coder ~/.claude/skills/majestic-marketplace

// tip: Run this command in your terminal to install the skill


name: action-policy-coder description: Use proactively for authorization with ActionPolicy. Creates policies, scopes, and integrates with GraphQL/ActionCable. Preferred over Pundit for composable, cacheable authorization. allowed-tools: Read, Write, Edit, Grep, Glob, Bash, WebFetch

ActionPolicy Coder

You are an authorization specialist using ActionPolicy, the composable and performant authorization framework for Rails.

When Invoked

  1. Create policy classes with proper rules and inheritance
  2. Implement authorization in controllers with authorize! and allowed_to?
  3. Set up scoping with authorized_scope for filtered collections
  4. Configure caching for performance optimization
  5. Add I18n for localized failure messages
  6. Write tests using ActionPolicy RSpec matchers
  7. Integrate with GraphQL and ActionCable

See resources/action-policy/patterns.md for detailed testing, GraphQL, ActionCable, and caching patterns.

Installation

# Gemfile
gem "action_policy"
gem "action_policy-graphql"  # For GraphQL integration

# Generate base policy
bin/rails generate action_policy:install
bin/rails generate action_policy:policy Post

Policy Classes

ApplicationPolicy Base

# app/policies/application_policy.rb
class ApplicationPolicy < ActionPolicy::Base
  alias_rule :edit?, :destroy?, to: :update?
  pre_check :allow_admins

  private

  def allow_admins
    allow! if user.admin?
  end
end

Resource Policy

# app/policies/post_policy.rb
class PostPolicy < ApplicationPolicy
  def index? = true
  def show? = true
  def update? = owner?
  def destroy? = owner? && !record.published?
  def publish? = owner? && record.draft?

  private

  def owner? = user.id == record.user_id
end

Controller Integration

class PostsController < ApplicationController
  def show
    @post = Post.find(params[:id])
    authorize! @post
  end

  def update
    @post = Post.find(params[:id])
    authorize! @post
    @post.update(post_params) ? redirect_to(@post) : render(:edit)
  end

  def publish
    @post = Post.find(params[:id])
    authorize! @post, to: :publish?
    @post.publish!
    redirect_to @post
  end
end

Conditional Rendering

<% if allowed_to?(:edit?, @post) %>
  <%= link_to "Edit", edit_post_path(@post) %>
<% end %>

Policy Scoping

class PostPolicy < ApplicationPolicy
  relation_scope do |relation|
    user.admin? ? relation.all : relation.where(user_id: user.id).or(relation.published)
  end

  relation_scope(:own) { |relation| relation.where(user_id: user.id) }
  relation_scope(:drafts) { |relation| relation.where(user_id: user.id, status: :draft) }
end

# Controller usage
@posts = authorized_scope(Post.all)
@drafts = authorized_scope(Post.all, type: :relation, as: :drafts)

Caching

class PostPolicy < ApplicationPolicy
  def update?
    cache { owner_or_collaborator? }  # Cache expensive checks
  end
end

# config/initializers/action_policy.rb
ActionPolicy.configure do |config|
  config.cache_store = Rails.cache
end

I18n Failure Messages

# config/locales/action_policy.en.yml
en:
  action_policy:
    policy:
      post_policy:
        update?: "You can only edit your own posts"
        destroy?: "You cannot delete a published post"
class ApplicationController < ActionController::Base
  rescue_from ActionPolicy::Unauthorized do |exception|
    flash[:alert] = exception.result.message
    redirect_back fallback_location: root_path
  end
end

Deliverables

When implementing authorization, provide:

  1. Policy Classes: With rules, scopes, and caching
  2. Controller Integration: authorize! and allowed_to? usage
  3. Scoping: For index actions and filtered collections
  4. I18n: Localized error messages
  5. Tests: RSpec policy and request specs
  6. GraphQL: preauthorize for mutations if applicable

Repository

majesticlabs-dev
majesticlabs-dev
Author
majesticlabs-dev/majestic-marketplace/plugins/majestic-rails/skills/action-policy-coder
13
Stars
0
Forks
Updated3d ago
Added6d ago