Why should I retry at the workflow level instead of inside functions?

Keeping resilience at the workflow level ensures business logic remains clean, testable, and focused on its primary task, while centralizing the error-handling strategy and preventing messy nested logic.

What is a 'retry explosion' and how does this skill prevent it?

A retry explosion occurs when multiple layers (API, service, and DB) all retry the same failure independently. This skill enforces the rule of retrying at only one level to avoid DDoS-ing your own infrastructure.

How does jitter help in production environments?

Jitter adds randomness to retry intervals, which prevents many instances from 'thundering herd'—hitting a recovering service at the exact same time and knocking it down again.

Can I use this skill for non-idempotent operations like payments?

The skill advises against retrying non-idempotent writes unless an idempotency key is used. This prevents dangerous side effects like double-charging a customer during a network timeout.

Workflow Resilience Patterns

Name: Workflow Resilience Patterns
Author: jagreehal

byjagreehal

0•

API Development

Implements robust retry, timeout, and circuit breaker patterns at the workflow level to ensure system reliability while keeping business logic clean.

This Claude Code skill provides a structured framework for building resilient applications by separating transient failure handling from core business logic. By utilizing the @jagreehal/workflow library, it enforces best practices such as jittered exponential backoff, mandatory timeouts, and idempotency checks. It helps developers avoid common pitfalls like 'retry storms' (double-retrying) and thundering herd problems, making it an essential tool for building production-grade backend services and distributed systems that must gracefully handle network hiccups and downstream service instability.

Key Features

010 GitHub stars

02Per-attempt and total operation timeout enforcement

03Smart transient error detection for selective retries

04Decoupled resilience logic from business functions

05Circuit breaker support for external service protection

06Configurable exponential backoff with mandatory jitter

Use Cases

01Stabilizing flaky third-party API integrations and webhooks

02Ensuring database read reliability during peak traffic periods

03Managing idempotent multi-step operations in microservices

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jagreehal/jagreehal-claude-skills resilience

For use in Claude.ai and ChatGPT

Download Skill