When I'm reading literature about proof-of-stake blockchain, I found that it seems there is no concept about confirmation blocks (which is used to confirm a posted transaction is indeed honestly generated and in bitcoin we need 6 confirmation blocks that is enough). However, my question is that why we don't need confirmation blocks in Proof-of-Stake blockchain like Proof-of-Work blockchain?
For a proof-of-work blockchain, say Bitcoin, the confirmation length is 6, just to make sure any adversary doesn't create any additional blocks and make a new fork that's longer. That's possible because all you need is more computation, and no proof of identity.
However, in proof-of-stake blockchain protocols, you often need to provide a new block with some kind of proof that you are eligible to propose the new block. This instantly creates a confirmed block because for a particular slot, another party cannot create a valid block and proof later in time.
Take a look at the Algorand protocol for more details.