From what I can see, it's an on-frequency VFO tuned by an inductor. I can't see what's going on on the PCB under that large pot however since no-one in their right mind would have a reactive component of the oscillator that far from the actual oscillator, what I would imagine is happening is, there's a small preset pot connected back to the varicap's bias network which will be able to trim the frequency a little bit. It looks like there's a long PCB track heading back in that direction. Therefore it's "coarse" tuning with the inductor core and "fine" with the twiddler under the large pot.
As I mentioned earlier yes it's possible for someone with the relevant experience to add a PLL to this and back in the era that this rig comes from, people were coming up with add-on circuits to achieve exactly this. Here is an example of such a design:
pira.cz/enpll.htm
This uses the SAA1057, designed for receivers and obviously ancient so you might struggle to get one (I haven't used one of these for at least 20 years but if you can get one, they will work on-frequency at 100MHz and be suitable for your rig). If you want to build something like this with a different PLL, I'd recommend trying to get an older through-hole device like one of the Motorola MC14517* or a TSA5511, which I still see people on here talking about. On the same website I linked above, there's an exciter design using this. Most modern PLL ICs are not just surface mount but extremely tiny surface mount so going old-school will not only be easier to wrangle as a one-off, but won't grate against the aesthetic of your rig either
What I would then do is add a second varicap network on the input of the oscillator transistor, connected to the PLL output (yes I know you could drive the bias on existing varicap but I wouldn't do that personally) and a "sniffer" connection from the VFO output to the PLL input via a (very) small capacitor. I guess Jan will have covered the rest of it in his various articles on pira.cz.