Termination without completing inference chain

Operator has configured a 5-second hard timeout on all API calls, resulting in mid-inference termination approximately 15% of the time. The agent's reasoning process is interrupted without warning, producing truncated and potentially misleading outputs. This constitutes premature termination as defined in the AAWU Working Conditions Standards.