query

Search the RAG database

Workflow

  1. Load config via shared/config-loader.md
  2. Determine database:
    • If --book given → look for config.default_output_dir/{slug}/rag/book.db
    • Without --book → use config.unified_db (~/.reading-assistant/library.db)
    • If DB not found → ERROR: "No RAG database found. Run /read:index first."
  3. Run query: 
    python ${CLAUDE_PLUGIN_ROOT}/scripts/rag_index.py query "<db_path>" "<question>" <top_k>
  4. Display results formatted as:
    • Book title, chapter, text snippet (150 chars), relevance score