Traditional macroeconomic models can have difficulty representing heterogeneous behavior, nonlinear interactions, and path-dependent adjustment. This study develops a stylized agent-based model (ABM) as a computational laboratory for examining macroeconomic policy transmission. The framework contains heterogeneous households and firms together with a fiscal authority and a central bank, and it is interpreted as a proof-of-concept Economic Digital Twin (EDT) rather than a real-time data-synchronized replica of a national economy. Three scenarios are considered: a baseline, a fiscal expansion, and a monetary easing. In the reported simulations, fiscal expansion increases mean output by approximately 2.28\%, whereas monetary easing raises mean output by about 0.34\%. Inflation remains close to the model's 3\% target, while unemployment changes very little across scenarios and remains unrealistically high at about 35.5\%. The latter feature is traced to the simplified labor-market block and the absence of explicit search, matching, and wage-bargaining mechanisms. The results should therefore be interpreted as qualitative evidence on model transmission channels rather than as empirically calibrated policy forecasts. The proposed framework provides a transparent policy sandbox that can be extended through empirical calibration, richer labor and financial sectors, and additional validation procedures.