Consider a closed system of constant volume $V$, constant pressure $P$, temperature $T$, and Gibbs energy $G$ that is in thermal and mechanical equilibrium with surroundings. It is filled with $N$ particles of one type.
We divide the system into two components: component 1 which has volume $V_1$, pressure $P$, temperature $T$, particle count $N_1$, chemical potential $\mu_1$, Gibbs energy $G_1$, and component 2 which has volume $V-V_1=V_2$, pressure $P$, temperature $T$, particle count $N-N_1=N_2$, chemical potential $\mu_2$, Gibbs energy $G-G_1=G_2$.
The total change in Gibbs energy for this two-component closed system is
$$dG=dG_1+dG_2=\mu_1 dN_1+\mu_2 dN_2=\mu_1 dN_1-\mu_2 dN_1=(\mu_1-\mu_2)dN_1$$
Suppose the system is initially not in chemical equilibrium $\mu_1>\mu_2$, it will decrease the particle count in component 1 $dN_1<0$ until $dG=0$.
However, I am not sure how $G_1$ and $G_2$ change as the system approaches chemical equilibrium. In terms of Gibbs energy, chemical potential is defined as
$$\mu=\left(\frac{\partial G}{\partial N}\right)_{T,P}=\frac{G}{N}$$
Since $G$ and $N$ are extensive quantities, it follows that $\mu$ is an intensive quantity. This means the value of $\mu$ is independent of particle count. For example, even if the particle count is doubled, $\mu$ stays the same.
So I am not sure how it is possible for $\mu_1$ to decrease as $N_1$ decreases if chemical potential is an intensive quantity.