Selective attention regulates the activation of working memory (WM) representations. Retro-cues, presented after memory sample stimuli have been stored, modulate these activation states by triggering shifts of attention to task-relevant samples. Here, we investigated whether the control of such attention shifts is modality-specific or shared across sensory modalities. Participants memorized bilateral tactile and visual sample stimuli before an auditory retro-cue indicated which visual and tactile stimuli had to be retained. Critically, these cued samples were located on the same side or opposite sides, thus requiring spatially congruent or incongruent attention shifts in tactile and visual WM. To track the attentional selection of retro-cued samples, tactile and visual contralateral delay activities (tCDA and CDA components) were measured. Clear evidence for spatial synergy effects from attention shifts in visual WM on concurrent shifts in tactile WM were observed: Tactile WM performance was impaired, and tCDA components triggered by retro-cues were strongly attenuated on opposite-sides relative to same-side trials. These spatial congruency effects were eliminated when cued attention shifts in tactile WM occurred in the absence of simultaneous shifts within visual WM. Results show that, in contrast to other modality-specific aspects of WM control, concurrent attentional selection processes within tactile and visual WM are mediated by shared supramodal control processes.