Defines how autonomous agents converge on a shared decision through calibration-weighted voting and reversibility-tiered thresholds.