Quickened True Strike for +20 1/round
The following trick works with the RAW but exploits several problems in the rules. There will be a section at the end detailing these issues in terms of what's broken and how everything fits together.
So, the basic idea is to sacrifice your swift action to give yourself +20 to hit with your first attack every round (and your second attack as well if you have a round to prepare or don't attack on a round for whatever reason) as a cantrip. This works extremely well on a sniper or similar character-- it works much less well on a character that intends to make more than two attacks per round, as an arcane trickster may well want to do.
+20, however, is a lot. It's enough that you can pretty much not worry about the rest of your attack bonus until very high levels. I mean, the Tarrasque has 40 AC so even with wizard BAB and just 2 Dex you are hitting that on a 10+ at level 17. And at earlier levels, the effects are even stronger. In games that allow some means to access at-will quickened true strike I can generally assume my PC's first attack will hit.
Now, how you get at-will quickened true strike is a little complicated. First, you take the Wayang Spellhunter trait and pick true strike. Then you take the Sacred Geometry feat (if it doesn't meet your bar for 'worth it', nothing does. Sacred Geometry is one of the best feats in the game by far). You pick 'Quicken Spell' and 'Umbral Spell' as your free prerequisite-bypassing metamagics. You also actually take Tenebrous Spell-- this is a feat tax we need to pay to engage the effects of our level-lowering trait. We also put max ranks in Knowledge(Engineering).
Using Sacred Geometry, we declare that we are casting a Quickened Umbral Tenebrous True Strike. Then we roll a number of 6-sided dice equal to our ranks in Knowledge (engineering)-- probably at least 5 since we're an Arcane Trickster, so we have a 79% chance of success which goes up to 97% at 6th level and is effectively 100% at 7th level or as soon as we take Calculating Mind (source). We stop the game for a while while we use math to make that roll equal one of the numbers we want and then show our math to the GM and they say "Yes, that is some math there, you can have free metamagic now", at which point our spell is cast as a swift action because it is Quickened and as a cantrip because it has been lowered to a level 0 slot (Umbral means Tenebrous is a +0 metamagic). We then have also inadvertently cast darkness on ourselves that goes away as soon as we attack but fortunately true strike also ignores concealment so that doesn't actually matter much even if we can't see through it unless your opponent can move as an immediate action or you need to carry the truestrike over to the next round or something. We make our attacks, the first of which gets +20 to hit, and then we repeat next round as normal. If we fail our check, which is very unlikely (~20% chance at level 5, <5% at level 6, <.01% at level 7), we lose our swift action and don't get our +20 to hit that round.
Is the resource cost worth it?
Yes, generally. While two feats and a trait and having to take a specific skill we might otherwise not take and eating our swift action is a very high cost-- particularly just for guaranteeing your first attack hits-- the side benefits, in this case, are pretty massive and mean that there's a good chance the actual opportunity cost here is more like one feat and your Wayang Spellhunter trait being on the true strike spell and your swift action. That's a pretty steep cost, but one that I think is somewhat less than the benefit of having +20 to hit on your first attack each round and ignoring concealment chances.
The side benefit of Sacred Geometry is you can quicken any spell you want up to 5th level for free all the time, pretty much. That's nuts. You can also apply the benefits of any metamagic you actually have to any spell for free in the same way, as long as your total adjustment wouldn't go over 9th level without Sacred Geometry. There's no limit to how often you can do this, and it is really, really good. Quicken is obviously the best metamagic to have selected for Sacred Geometry, because it can be applied to so many spells and nothing good needs you to actually have it as a prerequisite. There is, therefore, functionally no cost to having taken Quicken Spell with Sacred Geometry beyond that inherent in taking Sacred Geometry itself. Umbral spell is good in that it also has nothing good requiring it as a prerequisite, but its effect is hardly competitive with e.g. Maximize. As an arcane trickster, being able to create darkness wherever we want is kind of cool, but it's not exactly an optimal use of Sacred Geometry's prerequisite bypassing metamagic nonsense, especially since we also have to take its prerequisite anyways.
Knowledge (engineering) fuels Sacred Geometry. We need it for true strike, sure, but we also need it for every other quickened spell we cast, which means we probably make good use of this skill a lot. It's also got its own uses inherent in actually being a skill that exists, but Sacred Geometry is obviously the big one.
For a spellcaster, the opportunity to have a level 1 spell as a cantrip instead is a big deal. Because of this, Wayang Spellhunter is almost a guaranteed pick for traits. True Strike is absolutely not the easiest or best option here-- most spells have way easier methods of getting a +0 metamagic on them-- and there's a good chance if they weren't taking true strike this way your Trickster would take cure light wounds or moment of greatness or even blood money as a cantrip via simpler metamagic shenanigans instead. That means that this particular cost is a bit worse than costing just any trait, because it means you can't do any of that even with more traits. Essentially, this costs your regional trait, and that's a much bigger deal than e.g. spending a trait on a family trait.
Why use these options?
The answer is premised off of using at-will swift-action true strike. There are better ways of achieving the same results at higher levels (e.g. at-will limited wish via a sentient magic item) but this answer sets out to give an option that is extremely effective right away (i.e. at level 5) rather than one which e.g. can't be part of character creation until level 16.
There are four basic ways one can get access to an arbitrary 1st level spell as a cantrip in Pathfinder. One can use metamagic lowering via Wayang Spellhunter in combination with a +0 metamagic, one can try to convince the GM to let them create a custom race with such a spell at-will, one can try to convince the GM to let them create a custom magic item that gives at-will access to the spell, and one can create a sentient magic item that has a metamagic-ed and so higher-level version of the spell as an at-will power (sentient items can only grant 4th-7th level spells at-will, for some reason).
Two of these methods (custom items and custom races) require GM buy-in, and in the case of a custom race it gives you the spell as a spell-like ability which means no metamagic and so doesn't really work for us anyways (we'd need access to a 4th/5th level spell at-will, which isn't a thing via that method). The custom item system prices continual true strike at 8,000 gp. Such an item would give you a +20 bonus on each and every attack. This is by far the best option, but it is also literally the first example given in the SRD for circumstances in which the table is garbage and you should ignore the prices it gives-- the SRD then recommends 200,000 gp as a fair price for continual true strike, which is more money than you are probably starting with (and also more money than is probably worth the ability in a WBL game). A magic item that can cast a 4th level version of true strike on us at-will using its own action each round costs at least 57,000 gp (we can get one for 57K, but we might want it be more expensive and also have other abilities or otherwise be fancier)-- that's the price for a sentient magic item with an at-will 4th level spell and an at-will 0th level spell, the cheapest normal power. 57,000 gp is alright for a 12th-13th level character since the benefit is so large, but you might not now or ever be a 12th level character. This answer mostly assumes you are level 5, the minimum expected level to have a level of Arcane Trickster, and fully assumes you are near that level. If you are actually more like level 10 or higher, consider just buying or having a wizard craft such an item. GP is cheaper than other character resources, even in a WBL game.
If, however, you are in fact close to level 5, the Wayang Spellhunter method is the one that is best for you. Normally that means taking the trait with regards to a spell and then taking Merciful Spell (for even incidentally damage dealing spells) or some other more esoteric +0 metamagic feat so that you can apply it to the spell and end up with a 0th level spell. If nothing else works, you can almost always apply Fleeting Spell, but even that doesn't help us here because true strike lasts only 1 round.
When picking a +0 metamagic fails, there is a fallback strategy. The Tenebrous Spell feat can be added to any spell and costs a +1 adjustment normally, but on spells with certain descriptors that adjustment is waived; it becomes a +0 adjustment metamagic. The Umbral Spell feat gives that descriptor to any spell. In combination, then, the effects of the Umbral Spell feat (normally a +2 adjustment) with the actual Tenebrous Spell feat (now a +0 adjustment) results in a 0th level spell. We have to use this method because there are basically no published metamagic options specific to Divination spells, which is where we would be looking for a +0 adjustment. Since we need the effects of a metamagic without the actual metamagic (and since we definitely don't want to burn two feats when we could just get a single better one) we need Sacred Geometry.
What things in this are objectionable
The biggest one is Sacred Geometry. People whine about metamagic spell level lowering abuse, but that is only secondarily as bad here. Sacred Geometry is a massive problem.
First of all, most people really struggle to do the math fast enough to not bother other participants in the game. I find rapid arithmetic somewhat entertaining and consequently have a fair bit of experience doing that, so when we first introduced the feat to our play group it was the power level of the feat rather than the timing that was an issue, but as soon as another member of the group tried to use the feat in play it did not work. It slowed the game down to a screeching halt every time they cast a spell, to the point where 1) they stopped using it as much because it cost play time even though using it more often would have been more optimal and 2) we ended up having them tell me the numbers and with me doing the math instead, which was not great for a variety of reasons.
Because of this, unless you also enjoy and have significant practice in playing the particular sort of arithmetic mini-game Sacred Geometry calls for, you really need to use a program like this and have it ready for use at the table for the feat to be wieldly. Realistically, though, it would probably make more sense to just shortcut the minigame and roll d% against the probability of a possible match at that point. And that's sort of the problem with the feat-- it wants you to play a mini-game and to reward you for doing well and punish you for doing badly but the primary punishment for doing badly is that figuring out a viable solution takes extra time and that's kind of a dumb balancing mechanism-- having to sit around working on a group math problem is not what most players are signing up for when they roll up a Pathfinder character, and waiting for the party to do math is not exactly the best part of GMing, even if you like math. So that's bad, and leads to the feat being banned a lot. But there's more.
Sacred Geometry is also a ridiculously overpowered feat. Like, you don't have to meet the prerequisites for the metamagic feats you pick to use it with. That's absurd. Instead of taking a metamagic feat, I can take any two metamagic feats and ignore their prerequisites? That is just patently, obviously, clearly unbalanced. It's more than twice as good as any metamagic feat-- you get the same effect and you also get another, prerequisite-free feat. And then it also doesn't require any spell slot adjustment for those feats. And it negates the spell slot adjustment you would have had if you happened to take any other metamagic feats. I mean, this feat singlehandedly moves Heighten Spell from the trash pile of bad metamagic feats onto the 'things seriously worth considering' list. Actually, in general, it moves all the metamagic feats up from a pretty suboptimal place to top-tier stuff again. Or, well, their effects anyways-- you can take Sacred Geometry multiple times and you've little reason to ever take an actual metamagic feat instead.
So, Sacred Geometry is often banned for power reasons as well.
Beyond Sacred Geometry, metamagic abuse is the next most contentious/problematic area here. There used to be two traits that lowered spell level via metamagic shenanigans, but Metamagic Master, the other trait in question, got FAQ-errataed to no longer be able to lower spells' levels below their starting level. Wayang Spellhunter-- despite being brought up in the same forum contexts-- was never errataed in the same way, but the effect is almost exactly the same, to the point that the d20pfsrd people added an editorial sidebar to their version of the Wayang Spellhunter trait that deals with the topic. While the FAQ's language (this trait was never intended to...) actually makes it clearer that the spell-level lowering abuse does, in fact, work, it is nevertheless the case that being able to lower the level of spells is a huge deal and in particular the ability to get a 1st level spell as a cantrip at level 1 can completely shatter expectations for low-level low-to-mid-op games. The power level coupled with how strong the evidence is that this is not the way the rules are intended to work results in this particular use of the trait often being banned.