The complex and non-linear dependencies inherent in emerging financial markets necessitate advanced filtering and allocation techniques to ensure portfolio resilience. This study proposes a novel network-tuned asset allocation model for the PRIBUMI Bursa Malaysia Zakat Index by integrating the Triangulated Maximally Filtered Graph (TMFG) approach with Shapley value-based optimization. Unlike traditional Mean-Variance models that often fail under extreme market correlations, this framework leverages network topology to enhance stock selection and weight distribution. The methodology is structured into four distinct phases. First, we process daily closing prices of Pribumi-indexed stocks to compute log returns and construct a high-dimensional correlation matrix. Second, the TMFG algorithm is applied to filter the network, preserving the most significant hierarchical dependencies while removing noisy links. Third, we analyze the resulting graph using centrality measures specifically degree, closeness, and betweenness alongside a calculated Peripheral Index to identify assets with high diversification potential. Finally, the selection is fed into a Shapley value framework, a cooperative game theory method that assigns portfolio weights based on each stock's marginal contribution to the collective risk-adjusted performance. The findings indicate that the TMFG-filtered network effectively isolates core influential stocks from peripheral diversifiers, with the latter often providing superior stability during periods of market stress. By utilizing the Shapley value for final allocation, the model ensures a fair and mathematically rigorous distribution of weights that accounts for the multifaceted interactions between assets. This research provides a robust tool for institutional investors and policymakers to mitigate contagion risk and optimize returns within the Malaysian equity market, offering a sophisticated alternative to conventional diversification strategies.