invalid?

There are no random tests.

Fixed tests is missing "checkmate without capture" case.

There are too many essential parts of algebraic notation moves untested and not mentioned at all: where are castling, pawn promotion, en passant, disambiguation of pieces? The Wikipedia article you're linking has all of these, so it's clear that you haven't read any of the material you've linked in the description, which doesn't look good on you (we do expect kata author to know what they're doing before writing a kata).

(Also, in initial code, argument should not be chess_notation. You're passing in a move, so name it move or something.)

Examples are not specifications. The description should still be complete, correct, and consistent without any examples. This is obviously not the case here. Specification by example leads to people trying to read your mind, which people are notoriously not good at.

Also, a description should be self-contained. Just linking to a Wikipedia article is not enough; you should reproduce the relevant part(s) in the description. This can be short, and you can still send people to Wikipedia for additional information, but the essence should be in the description, not under a link. If I know chess, but not notation, I should be able to get all necessary information from the description.

A more strictly factual, formal if you will, description, without all the "trust me" and "don't worry" might be more readable. Breaking the fourth wall confuses me, and possibly other people.

JavaScript has placeholder kata code of twoOldestAges

No Sample Tests

https://www.codewars.com/kumite/65b06141b0783a4e06f5849f?sel=65b06209b0783a4e06f584ad

This should be a translation on any of the existing "Hello, John" kata - if you can find one that doesn't have a Java version already.

This duplicate should be unpublished.