Likely Ignorant tech proposal to this problem:

When a paper contains an “invalid reference”, flag that paper in a database. Expose a service that allows you to share a list of references and receive any flags and the reasons. You can then more carefully scrutinize the papers you’re citing that might have referencing issues.

Bonus: display an entire reference tree and show branches with issues. Not that an issue invalidates a branch, but it might bring into question the integrity of the branch, especially if the reference in question is so key to the claims.