Most existing research in the area of Dark Software Engineering (DSE) – how software processes and tools can be misused to deceive, manipulate, or exploit stakeholders – has mostly focused on user-facing dark patterns. Process-level deception, such as misreporting or gaming of metrics, remains underexplored. This study examines deceptive self-reporting behaviours emerging in educational software projects within a metric-driven development environment. We introduce a taxonomy of twelve process-level dark antipatterns – irregularities in planning, estimation, and time tracking – that signal attempts to appear compliant or productive. Using a forensic pipeline that analyses YouTrack logs from a Master’s-level Software Engineering course, we automatically detect and visualise these behaviours through “Antipattern Detection Cards”. Our findings show that all 18 analysed teams displayed at least three dark antipatterns, with “Delayed Time Tracking”, “Effort Underestimation”, and “Delayed Estimation” being the most prevalent. These patterns may stem from inexperience, although it is not possible to rule out strategic metric manipulation. The results illustrate how even educational software projects can reproduce dark process dynamics seen in industry, where quantitative metrics incentivize deceptive reporting. We discuss how automated process forensics can support early detection, ethical awareness, and prevention of metric gaming.