#artificial intelligence
May 2026
How Coding Agents Fail Their Users: A Large-Scale Analysis of Developer-Agent Misalignment in 20,574 Real-World Sessions
An observational study of 20,574 coding-agent sessions from 1,639 repositories across IDE and CLI workflows operationalizes misalignment as a breakdown made visible through developer pushback, and annotates each episode along four axes: form, cause, cost, and resolution.
Ningzhi Tang, Chaoran Chen, Gelei Xu et al.
· arXiv.org · 10 citations
· ⚡1