Which AI providers does this skill support?

The skill is provider-agnostic, recommending tools like OpenRouter for multi-provider access and Vercel AI SDK for flexible streaming and React integration.

Why is research mandatory for this skill?

LLM training data typically lags reality by months. This skill enforces using live tools like WebSearch and Gemini CLI to verify the latest model IDs, API syntax, and pricing before implementation.

How does /llm-gates help my project?

The /llm-gates command performs a quality audit across five pillars: routing, testing, observability, security, and cost, identifying gaps in your AI infrastructure.

What is Context Engineering in this context?

It is the philosophy that optimizing the entire LLM configuration, data flow, and environment (Context Engineering) is more effective than simply tweaking the wording of a message (Prompt Engineering).

AI & LLM Development Suite

Name: AI & LLM Development Suite
Author: phrazzld

byphrazzld

•

데이터 과학 및 ML

Standardizes the integration and optimization of Large Language Models through live research protocols and production-grade engineering patterns.

This skill provides a comprehensive framework for building LLM-powered applications, prioritizing real-time research over stale training data to ensure compatibility with the latest model versions and API changes. It guides developers through model selection using live benchmarks, implements optimized architecture patterns from simple calls to complex agents, and enforces best practices for prompt caching, structured outputs, and observability. By integrating tools like OpenRouter, Langfuse, and Promptfoo, it transforms LLM development from experimental prompting into a rigorous, production-ready engineering discipline.

주요 기능

01Quality Infrastructure Audits via the /llm-gates command

02Optimized patterns for RAG, Vector Storage, and Prompt Caching

03Mandatory Research-First Protocol to bypass AI training data lag

042 GitHub stars

05Standardized AI Stack integration for Vercel AI SDK and OpenRouter

06Model Selection Decision Trees based on live benchmarks like SWE-bench

사용 사례

01Building production-grade AI applications with reliable structured outputs

02Implementing observability and cost-tracking for high-volume LLM integrations

03Migrating between LLM providers or upgrading to the latest model versions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add phrazzld/claude-config ai-llm-development

For use in Claude.ai and ChatGPT

Download Skill