Even more promising was the nuclear lightbulb reactor [1], which was developed by United Aircraft Corporation (now UTC) for NASA under the Mars program in the 70s. They got just short of testing the engine with nuclear fuel and all of the remaining obstacles were with material science and computation which would have been very solvable given the tech developed independently in the 90s. The design could also have changed the trajectory of the entire nuclear power industry because it uses compressed uranium hexafluoride plasma to reach criticality so it has the benefit of using tens of kilograms of fuel instead of tens of thousands and it's an actively maintained magnetohydrodynamic system so if anything breaks or power is lost, the reactor slows down and fission stops. Large centrifuges are needed to recycle fuel from the buffer gas (when it's used as a terrestrial power plant instead of a rocket engine) so this reactor can also be self-breeding and recycle nuclear waste.
Many of the papers generated by UTC on this project were readily available a decade ago, though it seems many were reclassified since then.