We have the canonical function $\epsilon_p: \mathbb{Z} \to \mathbb{Z}_p, x \mapsto (\overline{x})_{k \in \mathbb{N}_0}=(\overline{x}, \overline{x}, \overline{x}, \dots )$. The function $\epsilon_p: \mathbb{Z} \to \mathbb{Z}_p$ is an embedding. We interpret with it $\mathbb{Z}$ as a subset of $\mathbb{Z}_p$.
Proof: Let $x \in \mathbb{Z}$ with $\epsilon_p(x)=(\overline{x})_k=0$, i.e. $x \equiv 0 \mod{p^{k+1}}$ for all $k \in \mathbb{N}_0$. The only integer number that is divisible by any high $p$-power is zero, so it holds $x=0$. Thus, $\epsilon_p$ is an embedding.
Could you explain me the above proof?
In general, to show a ring homomorphism $\varphi$ is injective, it is equivalent to show that $\varphi(0) = 0$, since $$\varphi(x) = \varphi(y) \iff \varphi(x-y)=0$$ so $\varphi$is injective if and only if $$x=y \iff \varphi(0) = 0$$ The proof has skipped the step of showing that your function $\epsilon_p$ is a ring homomorphism, and is just using the above fact to show it is injective.
We need to show that $\epsilon_p(x) = 0 \implies x = 0$. But by definition, $$\epsilon_p(x) = 0 \iff (\overline x, \overline x,\overline x,\ldots)= (0,0,0\ldots) \iff x \equiv 0 \pmod {p^{k}} \ \forall k \in \mathbb N$$ So $p^k \mid x$ for every $k$, and this can only occur if $x =0$.