AURA Agent Infers Implicit User Intent Before Tool Selection, +0.07 Coverage Gain

AURA inserts an IntentFrame inference step between scene perception and tool use, estimating implicit needs with a scalar gap score that controls probe budget. On a 100-query benchmark, it improves implicit-need coverage over ReAct-style probing by delta=+0.07 (p<10^-6), with gains attributed to gap calibration rather than answer memorization.