embedding-strategies

Quick Review

High-quality skill providing comprehensive guidance on embedding model selection and optimization for RAG applications. Excellent task knowledge with practical, production-ready code templates covering major embedding providers (Voyage AI, OpenAI, local models), sophisticated chunking strategies, and evaluation metrics. The description clearly indicates when to use this skill. Structure is clean with a logical progression from concepts to templates to best practices. The skill addresses a genuinely complex domain where naive approaches often fail—proper chunking, model selection, and evaluation require specialized knowledge that would consume many tokens for a CLI agent to discover independently. Minor points: While the skill is substantial, some aspects (like basic API calls) are somewhat standard, and the single-file structure, though well-organized, could potentially be modularized for a skill of this depth. Overall, this is a highly useful resource that meaningfully reduces the token cost and cognitive load for implementing robust embedding pipelines.