We assess the huff-n-puff performance in ultratight reservoirs (shales) by conducting large-scale numerical simulations for a wide range of reservoir fluid types (retrograde condensate, volatile oil, black oil) and different injection gases (CO2, C2H6, C3H8) by considering relative permeability hysteresis, diffusion, and sorption. A dual-porosity naturally fractured numerical compositional model is used that considers molecular diffusion and sorption to represent the flow mechanisms during the injection process. Killough's method, Langmuir's adsorption model, and Sigmund correlation are utilized to incorporate hysteresis, sorption, and diffusion, respectively. To investigate the impact of the fluid type, we consider three fluid types from Eagle Ford shale representing retrograde condensate, volatile oil, and black oil. We conduct a comprehensive evaluation of the impact of diffusion, sorption, and hysteresis on the production performance and retention of each fluid and injection gas. Eagle Ford formation is selected because it is the most actively developed shale, and it contains a wide span of PVT windows from dry gas to black oil. The simulation results show that the huff-n-puff process improves the oil recovery by 4-6% when 10% PV of gas is injected. The huff-n-puff efficiency increases with reducing gas-oil-ratio (GOR) as oil recovery from low (GOR) reservoirs is doubled, while recovery from retrograde condensate increased by 20%. C2H6 provides the highest recovery for the black and volatile oil, and CO2 provides the highest recovery for retrograde condensate fluid type. Diffusion and sorption are essential mechanisms to be considered when modeling gas injection to any fluid type in shales. However, the relative permeability hysteresis effect is not significant. Neglecting diffusion during the huff-n-puff process underestimates the oil recovery and retention capacity. The diffusion effect on the oil density reduction is observed more during the soaking period. The diffusion impact increases with higher GOR reservoirs, while the sorption impact decreases with higher GOR. The retention capacity of the injected gas decreases with higher GOR. The diffusion impact on the retention capacity increases with higher GOR. Hence sorption and diffusion must be considered when modeling the huff-n-puff process in ultratight reservoirs.