ask-rails 0.2.5 → 0.6.0

data/lib/ask/skills/rails.deploy_pipeline/SKILL.md

@@ -1,125 +0,0 @@
----
-name: rails.deploy_pipeline
-description: Pre-deployment checklist for Rails applications — migrations, assets, credentials, jobs, and logs
----
-
-Use this skill before or during a Rails deployment to ensure nothing is missed.
-Follow the steps in order — each builds on the previous.
-
-## Step 1: Check Pending Migrations
-
-Before deploying, always check for unapplied migrations:
-
-```ruby
-RunCommand.new.call(command: "bin/rails db:migrate:status | grep 'down'")
-```
-
-If there are pending migrations:
-1. Review them with `ReadFile.new.call(path: "db/migrate/<TIMESTAMP>_migration_name.rb")`
-2. Verify they're reversible: check `change`, `up`/`down`, or `reversible` blocks
-3. Check for data migrations (e.g., backfills) that should run separately
-4. Estimate execution time — large tables may need locking strategy
-
-## Step 2: Verify Assets Pipeline
-
-For Rails with assets (Sprockets or Propshaft):
-
-```ruby
-# Precompile locally to catch errors early
-RunCommand.new.call(command: "bin/rails assets:precompile 2>&1")
-```
-
-Check for:
-- Missing asset references (SCSS variables, image references)
-- JavaScript compilation errors
-- Asset fingerprint changes (manifold fingerprints if CSS changed)
-
-For importmap or esbuild/vite setups:
-
-```ruby
-# Check build config
-ReadFile.new.call(path: "package.json")
-ReadFile.new.call(path: "vite.config.ts") if File.exist?("Rails.root.join('vite.config.ts')")
-```
-
-## Step 3: Review Credentials and Secrets
-
-Verify that all required credentials exist in the target environment:
-
-```ruby
-# For production credentials
-RunCommand.new.call(command: "bin/rails credentials:show --environment production 2>&1")
-```
-
-Check for common secrets that may need updating:
-- `secret_key_base`
-- `database_password`
-- Third-party API keys (AWS, Stripe, SendGrid, etc.)
-- JWT signing keys
-- Any env vars referenced in `config/` files that aren't in credentials
-
-## Step 4: Check Background Jobs
-
-Review sidekiq/active job configuration before deploy:
-
-```ruby
-# Check job files for any that need queue configuration
-Glob.new.call(pattern: "app/jobs/**/*.rb")
-```
-
-For any new or modified jobs:
-1. Verify the queue adapter is configured in `production.rb`
-2. Check for job retry logic that might affect rollback
-3. Review `ReadFile.new.call(path: "config/sidekiq.yml")` if using Sidekiq
-4. Verify scheduled/cron jobs if using `sidekiq-cron` or `whenever`
-
-## Step 5: Review Production Log for Pre-deploy Issues
-
-Check that the current production environment is healthy before deploying:
-
-```ruby
-ReadLog.new.call(lines: 100, level: "ERROR")
-```
-
-If there are recent errors, investigate before deploying more changes.
-
-## Step 6: Verify Dependencies and Gem Versions
-
-Check for critical gem updates or changes:
-
-```ruby
-# Review Gemfile for new/modified gems
-RunCommand.new.call(command: "git diff HEAD -- Gemfile")
-```
-
-If adding a gem that needs system dependencies or native extensions:
-```ruby
-RunCommand.new.call(command: "bundle platform")
-```
-
-## Step 7: Config File Checklist
-
-Verify configuration files for the target environment:
-
-```ruby
-ReadFile.new.call(path: "config/environments/production.rb")
-ReadFile.new.call(path: "config/database.yml")
-ReadFile.new.call(path: "config/cable.yml") if File.exist?("config/cable.yml")
-ReadFile.new.call(path: "config/storage.yml") if File.exist?("config/storage.yml")
-```
-
-Key production checks:
-- `config.force_ssl = true`
-- `config.consider_all_requests_local = false`
-- Proper cache store configured (`:mem_cache_store`, `:redis_cache_store`)
-- Active Storage service configured for production
-
-## Step 8: Quick Rollback Checklist
-
-Before deploying, know how to roll back:
-
-1. **Database**: `RunCommand.new.call(command: "bin/rails db:migrate:down VERSION=<previous>")`
-2. **Code**: Git revert the deploy commit
-3. **Assets**: Previous version's assets should still be cached
-4. **Jobs**: Check if backward-incompatible changes won't replay failed jobs
-5. **Feature flags**: If using flipper or similar, toggle off new features first

data/lib/ask/skills/rails.route_trouble/SKILL.md

@@ -1,127 +0,0 @@
----
-name: rails.route_trouble
-description: Step-by-step methodology for debugging routing issues in Rails
----
-
-Use this skill when a route is returning 404, you're getting route matching errors,
-or routes aren't behaving as expected in a Rails application.
-
-## Step 1: Read the Routes File
-
-Start by examining the routes configuration:
-
-```ruby
-ReadRoutes.new.call
-```
-
-This reads `config/routes.rb` — the source of truth for all route definitions.
-
-Look for:
-- The overall structure (namespaced, nested, shallow routes?)
-- Resources that might be missing
-- Constraints or conditions that could block matching
-- Route ordering (more specific routes should come before less specific)
-
-## Step 2: Check the Compiled Routes
-
-Rails routes have a specific matching order. Use `RunCommand` to inspect the live routes:
-
-```ruby
-RunCommand.new.call(command: "bin/rails routes")
-```
-
-If you know the path you're trying to match, grep for it:
-
-```ruby
-RunCommand.new.call(command: "bin/rails routes | grep users")
-```
-
-For a specific route name:
-
-```ruby
-RunCommand.new.call(command: "bin/rails routes --grep user_profile")
-```
-
-## Step 3: Check Route Parameters and Constraints
-
-Routes often fail because of parameter mismatches or constraints. Check for:
-
-**Required parameters:** Does the route define `:id` but the URL doesn't include it?
-
-**Format constraints:** Routes with `:format` (like `.json`, `.html`) constraints
-can fail silently. Check if there's a default format:
-
-```ruby
-# In routes.rb:
-resources :posts, defaults: { format: :json }
-```
-
-**Constraint classes:** Custom route constraints like `subdomain` or request-based
-constraints can prevent matching:
-
-```ruby
-# In routes.rb:
-get "admin", to: "admin/dashboard#show", constraints: ->(req) { req.subdomain == "admin" }
-```
-
-Check the constraints against the actual request parameters.
-
-## Step 4: Trace the Match
-
-For a failing request, trace how Rails would match it:
-
-```ruby
-RunCommand.new.call(
-  command: "bin/rails runner \"puts Rails.application.routes.recognize_path('/users/1/edit', method: :get)\""
-)
-```
-
-This will raise `ActionController::RoutingError` if no route matches — the
-error message tells you what routes were tried before failing.
-
-If the route exists but the controller isn't found:
-
-```ruby
-ReadFile.new.call(path: "app/controllers/users_controller.rb")
-```
-
-## Step 5: Check Namespace and Module Nesting
-
-Routes in namespaced controllers fail when the controller file is in the wrong
-directory or the module is misnamed:
-
-```ruby
-# Route:
-namespace :admin do
-  resources :users
-end
-
-# Expects:
-# app/controllers/admin/users_controller.rb
-# class Admin::UsersController < ApplicationController
-```
-
-Use `Glob` to verify the controller exists:
-
-```ruby
-# Via RunCommand:
-RunCommand.new.call(command: "ls app/controllers/admin/")
-```
-
-## Step 6: Verify Helper Paths and Named Routes
-
-If you're debugging a `No route matches` error from a view (link_to, form_with):
-
-1. Check the route name: `RunCommand.new.call(command: "bin/rails routes | grep user_path")`
-2. Verify the route helper: `RunCommand.new.call(command: "bin/rails routes --grep user")`
-3. Check for route helper overrides in custom constraints or defaults
-
-## Failure Mode Guide
-
-| Symptom | Likely Cause | Action |
-|---------|-------------|--------|
-| 404 for existing route | Route constraint blocking | Check `constraints:` block in routes.rb |
-| `No route matches` | Missing resource definition | Add `resources :model_name` |
-| Route works in dev, not prod | Different route ordering | Check routes file isn't conditionally loaded |
-| `Missing controller` | Wrong namespace or filename | Verify file path matches module structure |
-| Path helper not responding | Wrong route name or params mismatch | Check `bin/rails routes --grep helper_name` |

data/lib/generators/ask/rails/install/install_generator.rb

@@ -1,29 +0,0 @@
-# frozen_string_literal: true
-
-require "rails/generators"
-
-module Ask
-  module Rails
-    module Generators
-      class InstallGenerator < ::Rails::Generators::Base
-        source_root File.expand_path("templates", __dir__)
-
-        desc "Creates ask-rails configuration and migration"
-
-        def create_initializer
-          template "initializer.rb", "config/initializers/ask_rails.rb"
-        end
-
-        def create_migration
-          timestamp = Time.now.utc.strftime("%Y%m%d%H%M%S")
-          template "migration.rb", "db/migrate/#{timestamp}_create_ask_sessions.rb"
-        end
-
-        def create_tools_directory
-          empty_directory "app/tools"
-          create_file "app/tools/.keep", ""
-        end
-      end
-    end
-  end
-end

data/lib/generators/ask/rails/install/templates/initializer.rb

@@ -1,5 +0,0 @@
-Ask::Rails.configure do |config|
-  config.default_model = ENV.fetch("ASK_DEFAULT_MODEL", "gpt-4o")
-  config.max_turns = ENV.fetch("ASK_MAX_TURNS", 25).to_i
-  # config.persistence_adapter = Ask::Rails::Persistence.new
-end

data/lib/generators/ask/rails/install/templates/migration.rb

@@ -1,12 +0,0 @@
-class CreateAskSessions < ActiveRecord::Migration[7.1]
-  def change
-    create_table :ask_sessions do |t|
-      t.string :session_id, null: false
-      t.string :model
-      t.jsonb :data, default: {}
-      t.timestamps
-    end
-
-    add_index :ask_sessions, :session_id, unique: true
-  end
-end