Distributed systems often need to perform several steps across different services. If one step fails, you need a way to undo the previous steps. This is where saga patterns help. They allow you to manage distributed transactions without complex locking.
This skill teaches you how to design compensating actions for each step. You can handle failures gracefully and keep your data consistent. It works with event-driven systems like Kafka or RabbitMQ.
You will learn to set up timeouts, state monitoring, and dead-letter queues to avoid stuck sagas. These practices make your long-running workflows reliable and easy to debug.
Global
mkdir -p ~/.claude/skills/saga-orchestrationProject
mkdir -p .claude/skills/saga-orchestrationSource Repository
Lark Openapi Explorerlarksuite/cli
Discover and call native Lark OpenAPIs not covered by existing CLI commands
Supabasesupabase/agent-skills
Add a database auth and APIs to your app with Supabase
Design An Interfacemattpocock/skills
Generate many different interface designs and pick the best one
Convex Quickstartget-convex/agent-skills
Rapidly set up a Convex backend for new or existing apps with one command
Convex Create Componentget-convex/agent-skills
Create reusable Convex components with isolated tables and clear APIs
Convexget-convex/agent-skills
Use this router to find the perfect Convex skill for your task
Firebase Remote Config Basicsfirebase/agent-skills
Change your app's behavior instantly without a new update
Notion Apiintellectronica/agent-skills
Automate your Notion workspace with simple API calls