Behavioral entropy is a Shannon-entropy measure computed over the timing distribution of interaction events. Human input carries irregularity that automation generally does not reproduce, so entropy is one of the discriminating signals in a session signature. Entropy alone is not a verdict: it is combined with linearity, hesitation structure, and micro-correction patterns, and the composite is compared against the session's established signature. MagenTrust stores the derived value, not the events it was computed from, and raw pointer coordinates and keystrokes are processed in memory and discarded.