How do you release all the resources created? I could not find any instances of `free()` in the code (and there are two calls to `malloc()`). Are you supposed to call `free()` yourself directly on the executor / task pointers? If so, I would suggest exposing a documented API to dispose of resources.
The tasks themselves come from a pool and are reused automatically, so no call to `free` is needed there. As for the executor, destroying it is a bit involved, because you need to wait for the currently executed task to finish before joining the threads and releasing the resources. It is a feature I do want to work on in the future though.
The library is an "stb-style" header-only library, so the header has two modes: when you define CR_TASK_IMPL, it includes the implementation (starting from line 20), otherwise it just includes declarations (the first 19 lines). The implementation should only be included in one translation unit, but the declarations can be included in multiple units.