I know this seems a bit obvious, but be sure the transistor still has gain to something like 2 or 5 times the highest frequency of operation.
It would help to know the frequency range you have in mind. The resonant structures that suit one band only mess up those for another. At HF, one has to switch component sets and gain-setting components and use several oscillators to seamlessly cover the range.
If by some chance the frequency range you require can be covered by a fast clocked DAC and pre-set sinewave accumulator, such as from a modern digital synthesizer chip. You need only a small low-pass filter after to remove the clock, and you can generate waves at will.
BUT.. you did say one transistor. Using the suggested 2-transistor push-pull method definitely has more potential, and was the basis for a HF gate-dip oscillator design that spanned several decades frequency. The resonator components and gains still had to be changed (using plug-ins) to cover the range ~1MHz through to 480MHz.