A structured workflow for efficiently exploring, mapping, and extracting information from 'rag-test' collections using RAG (Retrieval-Augmented Generation) and Knowledge Graph tools.
This skill provides a high-signal strategy for navigating and extracting knowledge from collections in the rag-test environment. It prioritizes synthetic knowledge (Summaries/Wiki) before deep-diving into raw text chunks.
Always start by identifying the exact collection name and its high-level structure.
collection_list() to find the exact name.get_summaries(collection_name, summary_type=2) (Type 2 is BOOK) to get genre, main characters, and themes immediately.get_table_of_contents(collection_name) to understand chapter progression and chunk distribution.Before using keyword searches on raw text, use the Wiki tools to get synthesized character and plot data.
get_wiki_index(collection_id) to see available characters, locations, and concepts.get_wiki_page(page_id) for specific entities (e.g., "Captain Ahab"). Wiki pages often contain curated relations and significance that raw chunks lack.Use the Knowledge Graph to verify structural and character relationships.
MATCH (n) RETURN labels(n), keys(n) LIMIT 5 if you are unsure of the graph structure.MATCH (c:COLLECTION {name: 'Name'})-[:CONTAINS]->(ch:CHAPTER) RETURN ch.nameMATCH (p1:PERSON {name: 'A'})-[r]-(p2:PERSON {name: 'B'}) RETURN type(r)Use query_collection only after you have a specific target or need direct quotes.
get_chunks_by_id to read surrounding context if a query result is truncated or needs more "flavor."wait_for_previous parameter with rag-test MCP tools unless explicitly confirmed to work; standard parallel calls are safer.MATCH ()-[r]->() RETURN DISTINCT type(r) to see what's available.get_chunks_by_id for what happened; use get_summaries or get_wiki_page for why it matters.