Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

Fix medium thinking chat failed in thinking (#18008)

Wang Qiba7d65a4ec25
diff --git a/rag/advanced_rag/harness/orchestrator/decompose.py b/rag/advanced_rag/harness/orchestrator/decompose.py
index 925dfb6d9b..0a5aaa23a6 100644
--- a/rag/advanced_rag/harness/orchestrator/decompose.py
+++ b/rag/advanced_rag/harness/orchestrator/decompose.py
@@ -146,7 +146,6 @@ async def decompose_and_search(state: dict, tools) -> dict:
gaps=analysis["gaps"],
discovered_claims=[],
)
- ctx.agent_results[c.claim_id] = c.agent_result
next_queries = _new_queries(
analysis.get("next_queries", []),