Submitted by Hengjie Zhu 201 Trial by Trace: Benchmarking AI Agent Harnesses for Reliability and Failure Attribution TraceLite AI 23 6