4 ms·
I think I have an explanation, but I'm not sure, so I would appreciate criticism. OP = person who posted pastebin E = person who's been stealing and launderin
by pxrdc 9y ago
I think I have an explanation, but I'm not sure, so I would appreciate criticism.
OP = person who posted pastebin
E = person who's been stealing and laundering funds
The key observation is that some people "create private keys from just about anything using Sha256 (i.e. Sha256(password/phrase)). This, of course, is NOT a recommended way of obtaining a private key since if YOU can think of the word/phrase, someone else can too"
Maybe E noticed this in 2014 just like OP did. Then she decided to try to make a lookup table. First she hashed everything in a natural language dictionary. Then to try to make her hash lookup table a bit bigger, extended it to any string of a given length like a transaction id, wallet public key, etc that appeared in the blockchain. This is somewhat similar to a standard cryptanalysis technique - where you try using everything in RAM as a key to see if it's the LUKS key - but using these particular inputs is admittedly ad hoc.
However, the next textbook step to attempting a lookup table is to do rainbow tables - which E started to do, but then gave up after a few iterations since the space was too large, and apparently this was enough, to have found the private keys for a significant number of other people's wallets.
Once the E had that, it's trivial to write a script that performs onward transfers to other accounts controlled she controls and steal money. It also makes sense to transfer the funds through several wallets as a laundry. It's strange that the transfers occur precisely along the links in the rainbow table (a rainbow table is really a forest of linked lists) - very strange. It's possible E just got lazy, since this is one of the easiest ways to write the script that decides where to send the money, once she controlled all these sha256-chains (chains in the rainbow table sense) already. Of course, that laundry's not good enough, since OP has now discovered it and made all of us curious to identify E.
- pxrdc 9y agoI would appreciate any criticism of this hypothesis. Far fetched as this is, I don't really see why OP assumes an "exploit," when good old fashioned precomputation - rainbow tables - explains the observation. Thoughts? I may be totally off. PS. There may be records of IP addresses involved in some of these transactions. Does anyone know where to find that kind of info, to see what IPs E was running the scripts from? People using VPSs, VPNs and Tor sometimes leak their home IP.