When to Use
Use this atomic skill for focused work on document-rag without bundling unrelated topics.
Quick Reference
This skill was split from vision-rag.md to keep topics independent and self-contained.
Core Patterns
Use the parent material below as the source reference, then keep implementations specific to this topic.
Source Reference (from merged skill)
name: vision-rag description: Vision-language models (VLMs), ColPali late-interaction retrieval for documents, RAG
[Description truncada. Veja o README completo no GitHub.]