The whole idea of a "born secret" novel propulsion technology seems unlikely to me. You need to start with basic research showing such a technology is possible in principle–and nobody would make that classified because it is way too remote from practical application. If you did, all that would happen is someone else would discover the same thing sooner or later, publish it openly, and then your own researchers would be demoralised about being denied credit for their own breakthroughs.

And then it is a very long and hard slog to get from basic research to something practically usable. Even if at some point you start classifying it, other teams can follow the same path openly in parallel.

Something is unlikely to be "born secret" unless it is developed in a government lab like LANL or LLNL. Are those kinds of labs even working on spacecraft propulsion? And even if they are, are they likely to be getting ahead of open research (academic, private sector, non-classified government labs)? Being classified can actually slow research down, by adding bureaucratic red tape and removing opportunities for collaboration and getting more eyeballs on a problem.

This sounds more like something out of fiction than reality.

The Manhattan Project was different because it was during a World War.