The Maritime Continent remains a long-standing challenge for numerical weather prediction (NWP). Accurate prediction of tropical convection over this region is further complicated by its small spatial scales, rapid evolution, and strong nonlinearity. Geostationary infrared (IR) satellite observations are widely regarded as one of the most valuable data sources for regional NWP by offering high temporal and spatial resolution over a broad domain. This capability enables near-continuous monitoring of rapidly evolving weather systems from mesoscale to convective scale. Therefore, this study investigates the impacts of all-sky IR radiance assimilation on cloud and precipitation forecasts over this area. Both water vapor channels 8-10 and the cloud-sensitive window channel 13 from the new-generation Himawari-9 Advanced Himawari Imager (AHI) are assimilated using hybrid 3D/4DEnVar methods within the MPAS-JEDI system. Cycling assimilation experiments are conducted to systematically evaluate their impacts on the analysis, background, and forecast fields using multiple independent observations. Results suggest that, relative to clear-sky assimilation, the analyses of brightness temperatures and cloud-top heights from the all-sky AHI assimilation experiments exhibit a better fit to the all-sky observations. Background verification indicates overall neutral-to-positive impacts, with particularly pronounced improvements in humidity. Furthermore, short-range cloud and precipitation forecast errors are generally reduced in the AHI assimilation experiments. Adding channel 13 further enhances rainfall forecast skill during the first 12 hours, whereas the 4DEnVar framework yields more sustained improvements at longer lead times. These results underscore the promise of all-sky AHI radiance assimilation for improving convection-permitting forecasts over the Maritime Continent.