Recent work supports the role of reasoning in third-party moral judgment of harm transgressions. The dynamics of the underlying cognitive processes supporting moral judgment is however poorly understood. In two preregistered experiments, we addressed this issue using a two-response paradigm. Participants were presented with moral scenarios twice: they had to provide their first judgment about an agent under both time pressure and interfering load, and were then asked to respond a second time at their own pace. In Experiment 1, participants (...) were harsher toward a malevolent agent at the second response, assigning more moral wrongness and punishment to an agent who either attempted to harm or harmed intentionally. Experiment 2 replicated the effect of intention on response change in a paradigm contrasting accidental to intentional harm scenarios. Participants were not only harsher toward intentional transgressors at the second response, but they were also less harsh toward accidental transgressors at the second response. We discuss the possibility that decoding overall intent and assigning moral judgment based on the presence or absence of a malevolent intent may be a relatively costly process. (shrink)