All these things happened, and it's up to textual critics to work out where and when and how and what.

What if ancient authors hijacked the name of famous authors to improve sales

This certainly happened, and works that were attributed to a famous author but now have that authenticity doubted are listed as dubia, and in some cases convincingly rejected, are often listed as (for example), Pseudo-Athanasius (I'm using theological examples, but the same is true for pre-Christian texts).

Slips by the scribes

Scribes did make errors. Our main way of attempting to fix this occurs where we have multiple manuscripts, and even better multiple manuscript traditions. Then we compare them, ask questions about which mss is older, more reliable, etc., ask which direction the error is more likely to have occurred in, and finally make a choice about which is more likely a better reading.

In the cases where it seems like there's been an error, but there's no divergence among manuscripts at that point, editors will sometimes suggest emendations based on why they think there might be a textual corruption here. These are much more speculative.

What if book dealers liked to improve their books (say a trader goes to Carthage with Plato, and adjusts the goods to local tastes, and what if the new text beat the original?)

This did not happen so much. But it's not impossible. In that case we simply end up with the text that we have. We can't construct texts for which we don't have evidence.

What if scribes wanted to try fixing errors in the manuscripts they copied

While scribal practices varied through antiquity, I'd say there was more of a tendency to try and preserve texts as they were than to fix them, though 'fixing' does occur. It's much like the case of 'slips' by scribes - scribes make mistakes, and change texts; scribes sometimes try to 'fix' perceived mistakes, and change texts. The procedure form our point of view is the same.

What if ancient libraries created incentives to fabricate new works?

This, not so much.

There's a limit to all textual work though - you can only go back as far as the texts go. That's just the evidence we have.

To speculate for a moment: if we constructed a text that is the same version of Plato's Republic that everyone has ever had, except Plato originally penned it differently, does it matter? The Republic that isn't Plato's is the one everyone read and is the one that influenced all of Western Philosophy - that's the text we have and deal with.

On the whole, though, our manuscripts appear mostly reliable, and historians do their best to produce the most reliable versions of them based on the data available.

Manuscript tradition varies based on the work. Some works have better tradition than others. There are accounts Solon and Peisistratus trying to alter Homer. Scribes did try to fix what they saw as errors; they also made their own. But we can trace these errors back and try to recreate the original work. In terms of Latin and Greek texts, the best example of this would be the Oxford Classical Texts. These offer not just the text but also include critical apparati that give you different readings on words. And in fact because of the problems with manuscripts, people still edit the texts today and offer their own readings.

As for some of the more insidious things you suggest, those seem more far-fetched. If someone tried to steal the name of another author, they would need to match their style. And people did talk and exchange ideas so it would be surprising if a new work of Vergil just showed up in the second century. Book dealers wouldn't really be an occupation. The Greeks and Romans wrote on papyrus scrolls that were extremely expensive. It would not be feasible for someone to edit or reproduce a regional version of a text.