To me that sounds like an intermittent throttle position sensor. For the problem your describing, I would look at that first.
Personal experience is that if you are going to have a connection problem on a harness,check the interface between the male and female pins. Its possible to get what's called fretting corrosion where the connectors meet. The plating on the connectors surfaces is supposed to prevent it. If the plating is thin and or gets worn off, a microscopic layer of corrosion forms.
Unplugging and replugging in the connector wipes it way and you get a lower resistance connection. The down side is without the plating, that the corrosion reforms. The upside is you bought yourself a fair amount of time.
This is also why connectors have a cycle rating, you only get so many connect disconnect cycles before you take the plating off. If I remember right, 40 is a typical number.
This is one place where a dielectric grease may help since it will protect the metal.
As far as "fixing" inside the crimp, even if you managed to loosen the oxidation, where is it going to go so that its not in the compressed zone where the connection is made?
Also, electronics flux residue can cause wire damage if used improperly. What happens to the acid in the flux when it combines with moisture?
Its not commonly known that electronic rosin fluxes are acidic. Not muriatic acid like plumbing still acid, just a milder form (plicatic acid and abietic acid).
Improperly would be excessive application and a failure to bring all of the flux up to its activation temperature. Worse will be if the flux is bridging dissimilar metals and there is moisture present.
No-clean rosin-based rework fluxes should not need any cleaning as long as the flux gets completely activated. Typically a flux gets activated with heat.
In a rework operation, the person could "pour" a lot of flux thinking that this helps soldering. Also, when using a solder tip and wire for soldering, flux surrounding the solder joint may not see enough heat and therefore may not activate completely.
This is a potential reliability concern as unactivated flux could cause dendritic growth and corrosion.
In short... (pun intended) put an ohm meter across your TPS and sweep it a few times to look for intermittent spots. An old school analog meter is great for that test.
