I’d counter argue that a test that is not communicated, reported, described or otherwise transmitted to another party is identical to it not happening, therefore one needs to tell “someone” (even if that is a private journal), and while in theory falsifability is possible solo, it increases the problem of induction, and science is, in essence, a language: a description of phenomena not the phenomena itself.