Wow, so easy to read and understand. The technical details are hidden behind well named functions and it is a pleasure to follow the lines. I tried a similar approach but couldn't achieve such a level of clarity. Bravo !

Many questions and issues come from the description on how to divide the message and for my part I spent a lot of time trying to figure out the rules and conditions. After many tries, I managed to write a somewhat naive implementation that worked. That was quite unexpected and I still can't find a good description of the division rules.

In my opinion, the whole problem comes from the fact that the description is solution-based whereas it should always be problem-based. It is as if the author tries to describe the division from the algorithm without wanting to give too much hints on the algorithm itself. The rules should have a meaning in the context of the problem we're trying to solve. Do you imagine using such division rules in the real world, say in Ancient Egypt ?

A better approach would be to have division rules understandable from the context itself. It seems hard to do it from the current rules (what is the logic behind having a runner carrying only one character?) but I can give an example of other rules that could be described contextually.

1. The message must be divided between all five runners (they are paid to run after all)
2. No runner should have too few or too much information, that is the length of the five messages should be as close as possible
3. By convention, the parts with more information are given to the first runners

Applying these rules will of course fail on most of the current test cases and I don't necessarily suggest to rewrite the kata with them. I just wanted to take this kata as an opportunity to emphasize on the necessity of having a very clear and above all problem-oriented description.

The description is not completely clear on how \ and / should take into account non-commands. The description says that these two commands will skip the next command and I understand also that characters others than the 8 commands are not considered to be commands ("The program is also supposed to ignore non-commands").

For tape \h*&, the first command \ will skip the next command, that is * since h is not a command. Should not the result be an empty string (the unique * command has been skipped) in this case ?

This kata is a very nice entry point to the world of esolang and the brainfuck language. At the end I spent more time navigating through the links than solving the kata. Thank you.