REASON – Retrieval-enhanced Evidence Assessment with Synthesis and Organized Narration: An End-to-end agent-based LLM pipeline for evidence synthesis using retrieval, classification, and structured summarization.