If you look at GTOWizard you will notice the hand is mixed. This means that with optimal play KJo is 0EV for raise, call or fold.

The frequency of which BB raises is ~14%. If you notice that the player pool doesn't do this as often you have to ask yourself if the difference is in strong hands or bluffs. And what that means for the EV of a hand with different lines.
If BB is more value heavy then KJo's EV of call (or raise) will lower. It will lower to below 0 and your preference then will be to always fold KJo.
If it's more bluff heavy then your EV of call and raise will go above 0 and you will never fold.

I would say your initial idea of pure folding is a fine adaptation.

The proces to come to the conclusion can be multiple ways. One I explained in the previous paragraph. Understand the EV of a line and where that EV comes from. KJo will benefit from you hitting a K or J and your opponent bluffing with his bluffs, especially on K high boards. If there are less bluffs then that part drops. You'll still be dominated by better Kx, Jx, overpairs, or folding when you miss despite having some equity.

Another way is what you described and I did in my Equity Realization video. Run a database solve with the ranges, see what the average EV is off the hand compared to the money you have to put in. It's a 3bet pot so should have run in a few hours.

I'm not sure if there is a great way to do this in pokertracker. The hand range visualizer seems good but I think you can only get that to work on your own range and not your opponents.

You would have to do this by creating an alias of a lot of players. Then use the range visualiser for that alias.
It's much easier with Hand2Note's Range Research feature. For that feature you do need a subscription, which is monthly. It's a great tool and worth it if you want to delve into pool imbalances. You do need a large sample of hands on players for this to be worth it though.

You will need to find a program that is able to output the specific EV of each hand in your range and go from there. I use Snowie for this exact purpose, I don't know what other programs you can do this with. Snowie has predetermined ranges though so its not customizable. If you find a program that lets you do this with customized villain and hero ranges and sizes then it will be perfect.

Once you've done the tedious work of listing out the hands and their EV, you will see which segments of hands fall into the bottom of your range vs optimal 3bets, then fold these versus opponents that dont 3bet optimally. And if they are 3betting even tighter, you find the next segment of bottom range hands and fold those. Continue this process until you feel satisfied with your now stronger defending range. Now do that for EVERY SINGLE SPOT to achieve max autism.

Your intuition about KJo is correct in a pool of under 3betting players. Most solvers will continue with KJo on the BTN vs both BB/SB, and in SB vs BB 3bet. But in practice you can fold it since it is the bottom of your range vs optimal 3betters, it will be -ev vs tight 3bets.