diff --git a/mcp/tools/rag_retrieve_tools.json b/mcp/tools/rag_retrieve_tools.json index df750de..fa644b2 100644 --- a/mcp/tools/rag_retrieve_tools.json +++ b/mcp/tools/rag_retrieve_tools.json @@ -7,11 +7,11 @@ "properties": { "query": { "type": "string", - "description": "Retrieval query content. Optimize the query by keeping core keywords and removing vague modifiers. For complex questions, split into multiple sub-queries." + "description": "Retrieval query content. Before retrieval, rewrite the query to improve recall: extract the core entity, time scope, attributes, and intent; add meaningful variants such as synonyms, aliases, abbreviations, related titles, historical names, and category terms; expand enumeration-style queries more aggressively; preserve the original meaning and do not introduce unrelated topics; use both the original query and rewritten queries whenever possible. For historical or list-style queries, also add terms like title / organization variants, predecessor / successor, former / past / historical / all-time, and list / overview / roster / timeline / archive." }, "top_k": { "type": "integer", - "description": "Number of top results to retrieve (default: 100)", + "description": "Number of top results to retrieve. Use the smallest sufficient top_k and expand only when coverage is insufficient: 30 for simple fact lookup about one specific thing; 50 for moderate synthesis, comparison, summarization, or disambiguation; 100 for broad-recall queries needing high coverage, such as comprehensive analysis, scattered knowledge, multiple entities or periods, list / catalog / timeline / roster / overview requests, or all items / historical succession / many records. Raise top_k when query rewrite produces many useful keyword branches or when results are too few, repetitive, incomplete, sparse, or too narrow in coverage. Do not raise top_k just because the query is longer. Expansion sequence: 30 -> 50 -> 100. If uncertain, prefer passing 100. Default: 100.", "default": 100 } }, @@ -26,7 +26,7 @@ "properties": { "query": { "type": "string", - "description": "Retrieval query content for table data. Optimize the query by keeping core keywords and removing vague modifiers. For complex questions, split into multiple sub-queries." + "description": "Retrieval query content for table data. Before retrieval, rewrite the query to improve recall: extract the core entity, time scope, attributes, and intent; add meaningful variants such as synonyms, aliases, abbreviations, related titles, historical names, and category terms; expand enumeration-style queries more aggressively; preserve the original meaning and do not introduce unrelated topics; use both the original query and rewritten queries whenever possible. For historical or list-style queries, also add terms like title / organization variants, predecessor / successor, former / past / historical / all-time, and list / overview / roster / timeline / archive." } }, "required": ["query"]