Decoder Silencing: Cloned GUI Agents Know When to Ask, and Greedy Decoding Discards It

A GUI agent that sends messages, deletes files or makes payments should ask before acting on an instruction it cannot resolve. Policies cloned from demonstrations routinely ask on no such instruction, and the field reads a $0\%$ ask-rate as a behaviour the supervision failed to teach. We measure the probability such a policy puts on asking with one teacher-forced forward pass per held-out state, no rollouts and no outcome labels. Across $40$ reruns of a single recipe, with nothing varied that the trainer reads, it spans $0.0002$ to $0.9907$. Greedy decoding compares each against one half, so a run holding substantial ask-mass and one holding almost none deploy at the same $0\%$ ask-rate. We call the first case decoder silencing: the behaviour is present in the belief and discarded at the decoder, not missing from the policy. The distinction is actionable: only a silenced behaviour can be repaired without retraining, and the probe says before deployment which runs a decode-time threshold will repair. We test this on AmbiGUI, a testbed we release where ambiguity comes from environment state rather than instruction text and every instance is verified on both sides: the question resolves it, guessing causes an irreversible action. On a pre-registered set it repairs about half of them, cutting irreversible wrong actions and raising task success, though a minority of runs come out worse on each. On AndroidWorld the probe is blind before training, when the tiers do not separate, and names what the agent does after. There, a backbone trained on whole episodes and deployed live on the emulator asks on every ambiguous instance and finishes $79\%$ of them with no irreversible error. No arm completes a task when its demonstrations stop at the decision rather than the finished state. Put plainly, a policy that never asks has usually been silenced rather than left untaught — so a $0\%$ ask-rate should be measured before it is believed.