For example, if the code calls for copying a wallet.dat file and it's software for playing mp3's, you probably shouldn't be using it.
It's good for those who have the skills to detect such irregularities. Can you notice such things personally? There's so many ways that hackers can inject malicious code into software that it's really not a simple task to verify an open source software. A good example is a failed attempt to backdoor Linux kernel by changing a single character to introduce a privilege escalation bug - and it failed because it was done as a commit. If someone was reviewing the whole repo from scratch, it would be easy to overlook it.
There are also techniques for code obfuscation which would allow hackers to hide malicious code from searching for potentially dangerous code, like using file system, internet connection, etc.
If reviewing code was so easy, software development companies wouldn't have to hire as many programmers.
There are open source software that just uses some libraries which is already popular and tested by the others so knowing those would already count. Then, the main code of a software is often simple yet turns to have complex when you would read the objects and classes that contains the main logic and algorithm. Practice in coding and even code development in general is already a huge step towards not having a difficulty understanding a software's code. Also, checking it doesn't really means to know it line by line, as most codes already have their comments and documentation before being published to the public.
probably most of us do not even check all of the codes or sometimes we don't even visit it, personally, I don't even reach those lines, as well as not everyone could reach programming language.