name: local/basic-rag
description: >
  Build a basic RAG (Retrieval Augmented Generation) application in Python.
  Use OpenAI embeddings for vectorizing documents and FAISS for the vector store.
  Include a document loader that reads .txt files from a directory, chunks them,
  embeds them, stores in FAISS, and answers questions by retrieving relevant
  chunks and passing them to OpenAI GPT for answer generation. Include a CLI
  interface that accepts a question as input. Include requirements.txt.
strategy_actor: anthropic/claude-sonnet-4-20250514
execution_actor: anthropic/claude-sonnet-4-20250514
definition_of_done: >
  A rag.py implements document loading, chunking, embedding with OpenAI,
  FAISS vector storage, retrieval, and answer generation.
  CLI accepts a question via argparse and a --docs-dir flag for the documents directory.
  A requirements.txt lists all dependencies (openai, faiss-cpu, tiktoken, etc).