This is a collection of brainteaser puzzles that can be solved algorithmically. Each puzzle is explained in prose, and then a framework (possibly in a variety of languages) is provided that implements the bookkeeping of the puzzle. All you have to do is implement a single function that determines the actions taken at each step in the puzzle. Note that solutions ARE NOT provided in the repository.
The background behind this repo is a long discussion at work that occurred after one coworker presented the hundred prisoners puzzle (hence the repo name) to the rest of the group. Many people thought they had answers, but weren't able to express the solutions clearly enough to verify that they really worked in all cases. Since we're all programmers, I figured the best way to approach it would be to program a framework and then people could program their solutions. Then, it would become very easy to determine whether the solutions were 100% thought out (and, equally importantly, correct) -- and indeed it was!