Not without getting sued, no. But you could do a double blind on it. One person reads it and describes it in freeform prose, and another reads the prose then writes code to implement it as described. In the absence of documents describing the implementation details, this can sometimes work as it will avoid the copyright argument.

Still sketchy though, and you’d really have to verify isolation.