How does the skill handle API errors and rate limits?

It includes a 'safeLLMCall' utility that manages retries and implements exponential backoff specifically for 429 rate-limit errors.

Can I use this for complex reasoning tasks?

Absolutely. It includes Chain-of-Thought and Few-Shot prompting templates designed to improve the accuracy of model reasoning for complex queries.

Which LLM providers are supported by this skill?

The skill provides implementation patterns for major providers including OpenAI and Anthropic, covering both standard and streaming message completions.

Does this skill include RAG (Retrieval-Augmented Generation) logic?

Yes, it includes specialized patterns for document chunking, embedding generation, and querying vector databases like Supabase with pgvector.

LLM Application Development

Name: LLM Application Development
Author: julianromli

byjulianromli

•

101

Data Science & ML

Builds robust AI-powered features using standardized patterns for prompt engineering, RAG workflows, and multi-provider LLM integrations.

About

This skill provides a comprehensive toolkit for developers building Large Language Model applications, offering battle-tested patterns for prompt engineering (including structured, few-shot, and Chain-of-Thought techniques), seamless integration with OpenAI and Anthropic SDKs, and end-to-end RAG (Retrieval-Augmented Generation) implementations. It simplifies the complexities of document chunking, vector storage, and streaming responses while ensuring production readiness through built-in error handling and exponential backoff strategies for rate-limited APIs.

Key Features

Standardized SDK integrations for OpenAI and Anthropic providers
Complete RAG pipeline logic with document chunking and vector search
101 GitHub stars
Optimized streaming response patterns for low-latency user experiences
Advanced prompt engineering patterns including Few-shot and Chain-of-Thought
Resilient error handling with configurable retries and exponential backoff

Use Cases

Implementing automated text classification and sentiment analysis workflows
Building context-aware AI chatbots using proprietary data sources via RAG
Developing complex AI agents requiring multi-step reasoning and structured output

About

Key Features

Standardized SDK integrations for OpenAI and Anthropic providers
Complete RAG pipeline logic with document chunking and vector search
101 GitHub stars
Optimized streaming response patterns for low-latency user experiences
Advanced prompt engineering patterns including Few-shot and Chain-of-Thought
Resilient error handling with configurable retries and exponential backoff

Use Cases

Implementing automated text classification and sentiment analysis workflows
Building context-aware AI chatbots using proprietary data sources via RAG
Developing complex AI agents requiring multi-step reasoning and structured output