[FEATURE] RAG integration for document-grounded conversations

[jwesleye]

Feature Description

Retrieval-Augmented Generation - load documents into vector database for context.

Problem/Motivation

Agents need access to large document collections that won't fit in context window.

Proposed Solution

chat_loop myagent --rag ./docs/

# Automatically:
# - Chunks documents
# - Creates embeddings
# - Stores in vector DB (Chroma, Pinecone, etc.)
# - Retrieves relevant chunks per query

Benefits

• Large knowledge bases
• Accurate citations
• Up-to-date information
• Grounded responses

Priority

• Critical
• High
• Medium
• Low

[jwesleye]

Out of scope - RAG integration is beyond the scope of this chat interface