Uranium commonly occurs in an oxide form. The uranium ore in the Oklo natural reactors was U3O8, or yellow cake. Quite obviously yellowcake can serve as a reactor fuel. And there are real advantages to using a uranium oxide compound, rather than pure uranium metal as a reactor fuel. Pure uranium burns. So uranium dioxide, for example, would be a safer fuel form, but it is still undesirable for UO2 to come into contact with water in the presence of heat because of a small potential for corrosion. UO2 melts at something like 2800 C, so it can get very hot without melting. It is a poor thermal conduction, so he inside of uranium dioxide fuel pellets may become very hot during reactor operations. Once the heat heat from UO2 comes in contact with a zirconium cladding it is quickly transfered to the coolant. Because of its poor heat transfer qualities, UO2 fuel pellets are usually small.
UO2 fuel pellets are formed by heating UO2 power until it fuses into a ceramic. Incased in zirconium cladding the UO2 ceramic has a number of desirable properties. It is unlikely to release fission products under all but very unusual circumstances.
Since UO2 is relatively stable from chemical and physical viewpoint something has to go very wrong with a reactor, before it poses a problem. Thus potentially dangerous radioactive materials are locked in a double barrier inside the fuel pellet. The first barrier is the Uranium Dioxide ceramic which locks the FP into a rock like structure. How good is the protection that the UO2 inner fuel pellet provides? Lets go back to the Oklo reactors for a check. The Oklo reactors formed in A rock like formation containing U3O8 ore. That ore was not entirely solid, in fact it had to be water permeable in order for the natural reactor to become critical. What happened to all of those radioactive materials that were produced in the 17 natural reactors? Surprisingly enough they stayed exactly where they were produced 1.7 billion years ago. Geologist checked it all, and came to the conclusion that the radioactive materials produced by the Oklo reactors had move dvery little during the last 1.7 billion years. The radio-active byproducts produced b the Oklo Reactors were the same byproducts produced in modern reactors. They were not moved by the ground water that repeatedly flowed through the OKlo ore body. Nor were they transported by other geological forces. Thus the UO2 pellet is a significant barrier to the release of radiation from a reactor. The zirconium cladding of the pellet forms a second barrier to isotope escape.
Thus very unusual conditions would be required to trigger an escape of radioactive materials from a UO2 fuel pellet.
A note on Naval reactor fuel: I have been unable to determine the composition or the original Nautilus reactor fuel formula. The Nautilus was able to travel 66,000 miles with the original core, suggesting a low level or enrichment. Later the range between refulings was extended considerably. It is not clear if the fuel was uranium in metal form, UO2, or a Uranium alloy. Recent naval reactors use a Zirconium-Uranium alloy with the Uranium enriched to 93% U-235.