TSAssistant: A Human-in-the-Loop Agentic Framework for Automated Target Safety Assessment

TSAssistant is a newly developed human-in-the-loop multi-agent framework aimed at automating Target Safety Assessment (TSA) by utilizing specialized subagents for different TSA domains. It features a hierarchical instruction architecture that separates coordination from domain expertise, employs programmatic execution hooks for workflow enforcement, and includes an interactive refinement loop for expert review. This framework enhances scalability and reproducibility in TSA processes, providing high reproducibility and evidential grounding, which is crucial for practitioners looking to improve the accuracy and efficiency of safety assessments in therapeutic target evaluation.