This work describes a dataset containing high-density EEG (hd-EEG) and surface electromiography (sEMG) to capture neuromechanical responses during a reaching task with and without the assistance of an upper-limb exoskeleton. It was designed to explore electrophysiological biomarkers for assessing assistive technologies. Data were collected from 40 healthy participants performing 10 repetitions of three standardized reaching tasks. A custom-designed touch panel was built to standardize and simulate natural upper-limb movements relevant to daily activities. The dataset is formatted according to the Brain Imaging Data Structure (BIDS) standard, in alignment with FAIR principles. To provide an overview of data quality, we present subject-level analyses of event-related spectral perturbation (ERSP), inter-trial coherence (ITC), and event-related synchronization/desynchronization (ERS/ERD) for EEG, along with time- and frequency- domain decomposition for EMG. Beyond providing a methodology for evaluating assistive technologies, this dataset can be used for biosignal processing research, particularly for artifact removal and denoising techniques. It is also valuable for machine learning-based feature extraction, classification, and studying neuromechanical modulations during goal-oriented movements. Additionally, it can support research on human-robot interaction in non-clinical settings, hybrid brain-computer interfaces (BCIs) for robotic control and biomechanical modeling of upper-limb movements.