The "selector" variable that the function takes is not a selector, but the name of the tag. Since find_all is used

find_all - does not support standard types of selectors (CSS, Xpath) and for filtering elements, it requires specifying target attributes using named variables or attrs. Docs

Example for "a.navbar__item":

soup.find_all("a", class_="navbar__item")
soup.find_all("a", attrs={"class": "navbar__item"})

For correct work with CSS selector, it is necessary to use - select. Docs

soup.select("a.navbar__item")

Therefore, with the current implementation of the function it is impossible to set a selector providing more accurate filtering of links