thibautbaissac/rails_ai_agents/caching-patterns

Claude Code and agent skill by thibautbaissac/rails_ai_agents

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Implements HTTP caching with ETags, fragment caching, Russian doll caching, and Solid Cache configuration. Use when optimizing performance, adding caching layers, or when user mentions ETags, fresh_when, stale?, cache keys, or Russian doll caching. WHEN NOT: For Turbo Stream real-time updates (use turbo-patterns), for background job cache warming logic (use job-patterns).

Install

npx vskill@latest install thibautbaissac/rails_ai_agents/caching-patterns

SKILL.md

Caching Patterns

Philosophy: Cache Aggressively, Invalidate Precisely

  • HTTP caching with ETags and fresh_when for free 304 Not Modified responses
  • Russian doll caching with touch: true for automatic cache invalidation
  • Fragment caching in views with cache keys based on updated_at timestamps
  • Solid Cache (database-backed, no Redis) for production caching
  • Collection caching with cache_collection for lists
  • Low-level caching with Rails.cache.fetch for expensive computations

Project Knowledge

Stack: Solid Cache (database-backed), Turbo for page refreshes, ETags with conditional GET, fragment caching in ERB views, collection caching for lists.

Multi-tenancy: Cache keys scoped to account. URL-based: app.myapp.com/123/projects/456.

Commands:

bash
rails solid_cache:install # Install Solid Cache
rails db:migrate # Run cache migrations
rails cache:clear # Clear cache

Caching Strategy Hierarchy

Apply caching in this order (highest impact first):

  1. HTTP caching -- fresh_when / stale? in controllers (free 304s)
  2. Fragment caching -- cache blocks in views (Russian doll)
  3. Collection caching -- cache_collection for lists of partials
  4. Low-level caching -- Rails.cache.fetch for expensive computations

Pattern 1: HTTP Caching with ETags

See @references/http-caching.md for full details.

ruby
# Single resource -- returns 304 if ETag matches
class BoardsController < ApplicationController
def show
@board = Current.account.boards.find(params[:id])
fresh_when @board
end
def index
@boards = Current.account.boards.includes(:creator)
fresh_when @boards
end
end
# Composite ETag from multiple objects
def show
fresh_when [@board, @card, Current.user]
end
# API with stale? for conditional rendering
def show
@board = Current.account.boards.find(params[:id])
if stale?(@board)
render json: @board
end
end
# Custom ETag with parameters
fresh_when etag: [@activities, @report_date, Current.user.timezone]

Pattern 2: Russian Doll Caching

See @references/fragment-caching.md for full details.

Set up touch cascades in models:

ruby
class Card < ApplicationRecord
belongs_to :board, touch: true
end
class Comment < ApplicationRecord
belongs_to :card, touch: true
# Updating comment touches card -> touches board -> invalidates all caches
end

Nest cache blocks in views:

erb
<% cache @board do %>
<h1><%= @board.name %></h1>
<% @board.columns.each do |column| %>
<% cache column do %>
<% column.cards.each do |card| %>
<% cache card do %>
<%= render card %>
<% end %>
<% end %>
<% end %>
<% end %>
<% end %>

Pattern 3: Collection Caching

erb
<%# Cache each item individually with multi-fetch optimization %>
<% cache_collection @boards, partial: "boards/board" %>
<%# Manual alternative %>
<% @boards.each do |board| %>
<% cache board do %>
<%= render "boards/board", board: board %>
<% end %>
<% end %>

Use counter caches to avoid N+1 in cache keys:

ruby
class Card < ApplicationRecord
belongs_to :board, counter_cache: true, touch: true
end
class Board < ApplicationRecord
def cache_key_with_version
"#{cache_key}/cards-#{cards_count}-#{updated_at.to_i}"
end
end

Pattern 4: Fragment Caching with Custom Keys

erb
<%# Multiple dependencies %>
<% cache ["board_header", @board, Current.user] do %>
<h1><%= @board.name %></h1>
<% if Current.user.can_edit?(@board) %>
<%= link_to "Edit", edit_board_path(@board) %>
<% end %>
<% end %>
<%# With expiration %>
<% cache ["board_stats", @board], expires_in: 15.minutes do %>
<div class="stats"><%= @board.cards.count %> cards</div>
<% end %>
<%# Conditional caching %>
<% cache_if @enable_caching, board do %>
<%= board.name %>
<% end %>
<%# Multi-key with locale %>
<% cache ["dashboard", Current.account, Current.user,
@boards.maximum(:updated_at), I18n.locale] do %>
<%= render "boards_summary", boards: @boards %>
<% end %>

Pattern 5: Low-Level Caching

ruby
class Board < ApplicationRecord
def statistics
Rails.cache.fetch([self, "statistics"], expires_in: 1.hour) do
{
total_cards: cards.count,
completed_cards: cards.joins(:closure).count,
total_comments: cards.joins(:comments).count
}
end
end
# Race condition protection for expensive operations
def expensive_calculation
Rails.cache.fetch(
[self, "expensive_calculation"],
expires_in: 1.hour,
race_condition_ttl: 10.seconds
) { calculate_complex_metrics }
end
# Version-based cache busting
STATS_VERSION = 2
def versioned_statistics
Rails.cache.fetch([self, "statistics", "v#{STATS_VERSION}"],
expires_in: 1.hour) { calculate_statistics }
end
end

Pattern 6: Cache Invalidation

See @references/cache-invalidation.md for full details.

ruby
# Prefer touch: true cascades (automatic)
belongs_to :board, touch: true
# Manual invalidation for low-level caches
class Card < ApplicationRecord
after_create_commit :clear_board_caches
after_destroy_commit :clear_board_caches
private
def clear_board_caches
Rails.cache.delete([board, "statistics"])
Rails.cache.delete([board, "card_distribution"])
end
end
# Sweeper pattern for batch invalidation
class CacheSweeper
def self.clear_board_caches(board)
Rails.cache.delete([board, "statistics"])
Rails.cache.delete([board, "card_distribution"])
Rails.cache.delete([board, "activity_summary", Date.current])
end
end

Pattern 7: Solid Cache Configuration

ruby
# config/environments/production.rb
config.cache_store = :solid_cache_store
# config/environments/development.rb
config.cache_store = :memory_store, { size: 64.megabytes }
# config/environments/test.rb
config.cache_store = :null_store

Pattern 8: Cache Warming

ruby
class CacheWarmerJob < ApplicationJob
queue_as :low_priority
def perform(account)
account.boards.find_each do |board|
board.statistics
board.card_distribution
end
end
end
# config/recurring.yml
cache:
daily_refresh:
class: DailyCacheRefreshJob
schedule: every day at 3am
queue: low_priority

Boundaries

Always
  • Use fresh_when for index and show actions
  • Use touch: true on associations for automatic invalidation
  • Use Solid Cache in production (database-backed, no Redis)
  • Include expires_in for time-based data
  • Scope cache keys to account in multi-tenant apps
  • Use counter caches for counts
  • Eager load associations to prevent N+1 queries
Ask First
  • Whether to cache user-specific content
  • Cache expiration times (freshness vs performance)
  • Whether to warm caches in background jobs
Never
  • Use Redis for caching (use Solid Cache)
  • Cache without considering invalidation strategy
  • Forget touch: true with Russian doll caching
  • Cache CSRF tokens or sensitive user data
  • Use generic cache keys without version/timestamp
  • Cache in test environment (use :null_store)
  • Cache across account boundaries in multi-tenant apps

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
589
Repository forks
67
7-day trend score
32

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/thibautbaissac/rails_ai_agents/caching-patterns/badge)](https://verified-skill.com/skills/thibautbaissac/rails_ai_agents/caching-patterns)

Something wrong with this skill or its page? Tell us.