I am indeed testing if a solution is reused rather than hard coding. Although it wouldn't be a nice solution people could hardcode the solution so I changed the description so reflect this notion.

Good suggestion and would make sense from methodical point of view but for a game it would be unfair if one player gets more cards than the other. So I included some tests that give one player slightly more cards than the other.

Thanks for having a look at the kata but I fear I cannot follow. There is a test that checks for matrixes overlapping. So can't be an 'R' in R's matrix when there is a 'B' in B's matrix. Hence, the matrixes are already required to be different.

That's great. I tried to model the rules of Codenames Duet in the kata but I don't know if there should be a "Black 8 Agent" like in the normal codenames version.

Thanks, I also changed the test a minute ago to be wrapped in a closure. Are we now in a limbo state with multiple valid katas?

Very bad practice to assign a lambda function to a variable. It works but is highly confusing.