Probabilistic number generation
I had this idea the other day but am wondering how to generate a wordlist/pipe it into hashcat.

Let's say you have a series of hashes, all of which have a 10 digit number as the input. Assuming the numbers are randomly generated, the probability of a single digit being 0-9 is 1/10. The probability of two of the same digits next to each other is 1/100, 3 digits is 1/1i000, and so on.

I first though permute could do this. While it generates all the permutations of 1234567890 just fine, there are duplicates when repeating one of the numbers(eg. 1224567890). If all 90+1 permutations for no repeating digits+1 repeating digits are calculated, that's 3.3 billion, clearly far too many duplicates.

My question is basically, any way to generate numbers where the least repeated digits are generated first?
anything wrong with just using the markov generator?
The markov generator is inadequate as it looks for similarities in the input. I do not want this.

Furthermore, it works by selecting the first number and then going through the entire keyspace minus the first digit. Again I do not want this. A number like 1000000 is totally unlikely and should be tried last.
you can take a look at the maskprocessor -q option, i think it's partially what you want to do. its not completely what you are looking for but maybe you can use it inside a script.
After some thought, I have come to the realization that maskprocessor with the -q option is the most effective for my purposes. Funny as I wrote a tool 2 weeks ago that does the same thing.

What's still left is sorting the output. But that's a challenge for another time for me.

Thanks for the tip.

edit: As I don't see the -q option in hashcat, are there plans to implement it?
No but you can use a fifo to pipe it