What are LLM Application Patterns?

These are standardized, industry-proven architectural blueprints for building complex AI software, covering everything from data retrieval to autonomous agent behavior.

Can I build multi-agent systems with this skill?

Yes, it includes patterns for multi-agent collaboration where specialized agents (researchers, analysts, writers) work together under a coordinator.

How does this skill improve RAG implementations?

It provides specific patterns for semantic chunking, hybrid search strategies, and reciprocal rank fusion to ensure more accurate and relevant context for your models.

Does it support prompt management?

Absolutely. It features patterns for prompt templating, versioning, and A/B testing to help manage the lifecycle of your prompts in production.

LLM Application Patterns

Name: LLM Application Patterns
Author: claudiodearaujo

byclaudiodearaujo

•

データサイエンスとML

Implements production-ready architectures for LLM applications, including RAG pipelines, autonomous agent patterns, and prompt management workflows.

LLM Application Patterns is a comprehensive architectural guide designed to help developers move beyond simple chat completions into building sophisticated, production-grade AI systems. This skill provides standardized implementations for Retrieval-Augmented Generation (RAG) with advanced chunking and hybrid search, diverse agentic frameworks such as ReAct and multi-agent collaboration, and robust prompt IDE patterns for versioning and evaluation. It serves as an essential blueprint for anyone designing AI-driven software that requires reliability, observability, and complex tool-use capabilities.

主な機能

01Advanced prompt engineering patterns for templating, versioning, and A/B testing

02Integration guidance for popular vector databases and embedding models

03End-to-end RAG pipeline patterns including semantic chunking and hybrid retrieval

041 GitHub stars

05Multi-agent collaboration frameworks for complex task decomposition

06Comprehensive agent architectures: ReAct, Function Calling, and Plan-and-Execute

ユースケース

01Establishing a professional LLMOps workflow for prompt versioning and observability

02Designing context-aware knowledge bases using advanced RAG and document-aware chunking

03Building autonomous research or coding agents that utilize external tools and reasoning loops

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/sistema-de-narra-o-de-livro llm-app-patterns

For use in Claude.ai and ChatGPT

Download Skill