What is the benefit of the Circuit Breaker pattern?

It prevents a single failing external service from overwhelming your system by temporarily blocking requests to that service until it is healthy again.

What are Resilience Patterns in the context of Claude Code?

Resilience Patterns are specialized architectural strategies that help AI-generated systems handle errors and failures gracefully without crashing the entire application.

How does this skill handle LLM API failures?

It implements fallback chains that automatically switch to secondary models or cached responses if the primary LLM API fails or encounters rate limits.

Does this skill help with managing LLM costs?

Yes, it includes token budget management patterns that monitor context usage and apply strategies like summarization or truncation to stay within limits.

Are these patterns production-ready?

Yes, these are production-grade patterns designed for high-stakes distributed environments and complex AI agent workflows.

Resilience Patterns

Name: Resilience Patterns
Author: yonatangross

byyonatangross

•

API開発

Implements production-grade fault tolerance and resilience patterns for distributed systems and LLM-based workflows.

The Resilience Patterns skill provides Claude Code with the architectural expertise needed to build robust, fault-tolerant distributed systems. It facilitates the implementation of industry-standard patterns such as circuit breakers to prevent cascading failures, bulkhead isolation to partition resources, and intelligent retry strategies with exponential backoff. Designed for modern AI development, it includes specialized logic for LLM API resilience, including model fallback chains and token budget management, ensuring that your AI agents and backend services remain reliable under load or during third-party outages.

主な機能

01Token budget guarding to manage context limits and optimize costs

02Bulkhead isolation patterns to prevent resource exhaustion across system tiers

03Advanced retry logic featuring exponential backoff and randomized jitter

04Circuit Breaker implementation with configurable failure thresholds and recovery timeouts

05LLM-specific fallback chains for multi-model reliability and graceful degradation

0669 GitHub stars

ユースケース

01Protecting distributed microservices from cascading failures using circuit breakers

02Optimizing LLM integrations with automated fallbacks and smart error classification

03Developing high-availability multi-agent systems that handle API rate limits and model outages

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit resilience-patterns

For use in Claude.ai and ChatGPT

Download Skill