No, you can give the model same prompt and it will give you a similar compaction result. On the backend, that's precisely what happens. There's nothing else going on in that encrypted blob, it's just summary of what model responds with when prompted "summarize current state and upcoming work".