This study verifies that a consumer-grade five-channel EEG headset can, through a unified real-time decoding pipeline, deliver reliable directional control across different game genres. We introduce the EEG Motor Imagery Game Controller (EMIGC). Muse 2 signals undergo standard preprocessing and sliding-window segmentation before entering a lightweight CNN-LSTM that decodes three sustained states and six transitional states, which are mapped to left/neutral/right commands. The same model seamlessly drives three Unity prototypes—Hide and Seek, Snake, and a two-lane Rhythm game, while multimodal feedback enhances the user experience. An event-based logger and a Formula Score provide fine-grained performance analysis, and questionnaire results indicate positive usability and immersion. EMIGC runs stably in all prototypes, demonstrating the feasibility of consumer-grade EEG for real-time, cross-genre game control.