A norm-governed agent takes social norms into account in its practical reasoning. Such norms characterise its role within a specific organisational context. By adopting a role, the agent commits to fulfil and adhere to the social norms associated with that role. These commitments require the agent to act in a way that does not violate any of its prohibitions or obligations. In adopting different sets of norms, an agent may experience conflicts between these norms as well as inconsistencies between possible actions for fulfilling its obligations and its currently adopted set of norms. In order to resolve such problems, it must be informed about conflicts and inconsistencies. The NoA architecture for norm-governed agents implements a computationally efficient mechanism for identifying and indicating such problems - possible candidates for action are assigned a specific label that contains cross-referenced information of actions and norms. As actions are indicated as problematic and not simply filtered out, the agent can still choose to either act according to its norms or against them. The labelling mechanism presented in this paper is therefore a critical step towards enabling an agent to reason about norm violations - the agent becomes norm-autonomous.