Let $A$ be an abelian group and $B$ a subgroup of $A$. Suppose $\sigma: A \to A$ is such that $\sigma(B) \subseteq B$, then it induces a map $\tilde{\sigma}: A/B \to A/B$.
I am interested in finding simple counterexamples (or proving the corresponding results, although I believe counterexamples should exist) to the following questions:
Given $\tau \in \text{End}(B)$, does there exist $\sigma \in \text{End}(A)$ such that $\sigma|_B = \tau$?
Given $\tau \in \text{End}(A/B)$, does there exist $\sigma \in \text{End}(A)$ such that $\tilde{\sigma} = \tau$?
Given $\tau \in \text{Aut}(B)$, does there exist $\sigma \in \text{Aut}(A)$ such that $\sigma|_B = \tau$?
Given $\tau \in \text{Aut}(A/B)$, does there exist $\sigma \in \text{Aut}(A)$ such that $\tilde{\sigma} = \tau$?
A counterexample for Question 3 has already been constructed in this question, which is quite clever. It was after seeing this counterexample that I was prompted to ask about the other cases, since a simple search did not lead me to find counterexamples for the other three questions on MSE. I think this is a very natural question and should be asked by someone.
Specifically, I am not sure if this problem can be viewed from a "higher perspective" using homological algebra methods.