Rehabilitation can promote brain plasticity and improve motor control after central nervous system injuries. Our working model is that motor control is encoded using dynamic primitives: submovements, oscillations, and mechanical impedances. We hypothesize that therapies focusing on these primitives can achieve greater motor recovery. At the observational level, these primitives lead to discrete and rhythmic movements. Here, we propose two novel biomarkers to evaluate rhythmic and discrete movements in gait based on the feet forward position: the smoothness of their relative position, using the mean-squared jerk ratio (MSJR), to assess rhythmicity; and the angle between principal components of consecutive trajectories (dPCA), to detect discrete movements amidst rhythmic motion. We applied these methods to kinematic data collected with healthy individuals during experiments employing the MIT-Skywalker: level-ground walking at five speeds, with and without imposed ankle stiffness; walking at constant speed on ascending, descending, and laterally tilted slopes; and performing sidesteps. We found a decrease in MSJR as speed increases, related to increased rhythmicity, even with imposed stiffness. Rhythmicity seems unaffected by the terrain perturbations imposed. Finally, dPCA successfully detects sidesteps, discrete events amidst rhythmic movement. These biomarkers appear to accurately assess rhythmic and discrete movements during walking and can potentially improve clinical evaluation and rehabilitation of neurological patients.