ai-llm-development

What problem does it solve?

The rapidly changing AI/LLM landscape makes it hard to stay current, leading to outdated model choices, inefficient prompting, and costly architectures. This Skill provides modern best practices to build robust, cost-effective, and scalable AI features.

Core Features & Use Cases

• Current Model Selection: Teaches how to use leaderboards and tools like OpenRouter to always find the best, most current models for any task.
• Prompt & Context Engineering: Optimizes LLM interactions through prompt caching, structured outputs, and system prompts, reducing costs by up to 90%.
• Architecture Patterns & Stack Defaults: Guides you through the complexity ladder (single call to multi-agent) and recommends a battle-tested stack (Vercel AI SDK, OpenRouter, Postgres pgvector).
• Use Case: When starting a new AI feature, activate this Skill to automatically recommend the optimal LLM, suggest prompt engineering strategies for cost reduction, and outline a scalable architecture, saving weeks of research and experimentation.

Quick Start

Apply the ai-llm-development skill to recommend the best current LLM for a coding task, considering cost and speed.