Introduction. The rise of autonomous agents is reshaping teamwork, introducing new errors not seen in human-only teams. Existing team error models fall short in addressing these challenges. This paper extends these theories to examine how HAT errors align or differ from human-only errors and their implications for competency modeling. By identifying transferable principles and unique challenges, we provide a framework for categorizing human-autonomy team errors. Background. Human-autonomy teams (HATs) involve at least one human and one autonomous agent whose performance is shaped by unique errors and recovery biases. Coordination challenges and trust perceptions influence human responses to autonomy in more pronounced ways relative to human-only teams. Sasou and Reason (1999) defined team errors as human errors in group processes, emphasizing both error-making and recovery. Proposed Model and Implications. HAT errors stem from humans, autonomy, or their interactions, classified as independent (e.g., internal flaws) or dependent (e.g., faulty external information). Errors can be individual (human or autonomy) or shared (human-human, autonomy-autonomy, or human-autonomy). The error recovery process consists of multiple barriers—detection, indication, and correction—that influence whether HAT errors are resolved. These barriers are strengthened or weakened by factors unique to humans and autonomy, shaping the effectiveness of error resolution differently by teammate type. Conclusion. Understanding HAT errors requires revising traditional performance models. As HATs grow, developers must classify autonomy errors effectively. Interdisciplinary collaboration is crucial to refine training and enhance human-autonomy teamwork.