If you can casually write an algorithm to break a modern cryptographic hash in 10 guesses... I would like to know. Because then I have to decide if I want to be a very good friend of you, once you get rich, or if I want to stay as far away from you as possible once the state intelligence agencies come after you.
It is not in the slightest bit viable. You’re seeking to reverse a one-way hash function. Knowing the full hash does not help you to find the original password; password cracking algorithms don’t work by reversing the hash, but by trying zillions of passwords, following typical human password patterns to increase the probability of success, and possibly using rainbow tables as precalculated hashess, until they find something that matches.
But we don't need to reverse the one way hash. The goal is simply to find the original password and brute forcing is good enough.
The brute forcing algorithm doesn't care that you only have a partial hash. All that does is increase the chances of collisions. (Side note, rainbow tables might care, I'm not sure how suitable they are for wildcard hash matches)
For example, I burned 8 guesses and I got enough greens to give me 108 bits of the hash. You can scrape out a bit more entropy by processing the yellows and greys, but 108 bits is more than enough to identify the password with very little chance of collisions (the chance of collisions hits only hits 50% once you get to 17 character alphanumeric+symbol passwords).
You can then use the two remaining guesses to resolve any collisions and lock in the correct answer.
The goal is to find the original password, but you’re finding the hash. Finding the hash doesn’t help you in the slightest with finding a password that hashes to that.
Put another way: here, I’ll tell you the hash: DF50B84AFEE438987ECE1542A4D1BCAB4079215EF38C3C3CBB2F4A122886DF27. Now tell me the password. You have 0% chance of succeeding in your lifetime, to at least a dozen decimal places.
