Working on pic related
Have searched this keyspace up to length of 3 chars for the F5 seed
abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789 .,:;!?()-+*/[]{}@_><#~=^`'"&%$
space included
No hits - proceeding to length of 4 - will report back in a few days
I haz both. I started on the evil eye but I noticed the rate was way low on that image (in comparison to test images)… you may want to check yourself. Much faster against illumipepe
https://arxiv. org/pdf/1606.00519.pdf
Can't we just use/modify the existing hashcat code for that?
https://hashcat.net/wiki/doku.php?id=example_hashes
Sorry that's for the AES decryption portion… still, I think we could use the existing hashcat code for the SHA portion of PRNG. SHA1/256 on hashcat is stupid fast. Something like 600m hashes/s on my old ass card.
I know. Rather than shooting for one target hash, we try 1k passwords at once and run each serially with however many iterations required, in parallel. I don't see a problem here. I still think it can be modified to our purpose.
Is size the size of the decompressed bitmap? Or is it something else?
PS tells me that's about 303K for illumipepe.
Even if it's 1MB as you say, that's still 1500 instances of the image.
With my lame 1.5GB graphics card that's still almost 5K potential instances
The DCT coefficient list only gets computed once, correct? If so, we only need to push one copy of the data to the graphics card and we should be able to copy it as many times as we want, no? And if we manage to implement it all on the graphics card, then all we really care about getting back is the rate of attempts and the valid key, if any. And yes, I understand many images are larger but essentially it would work out to max available GPU mem divided by decompressed image size in terms of threads. I'm willing to bet that's still a fuckton more than we've got going currently.
Roger that.If there's one thing I'm certain of though, it's that we drastically need to speed things up. Perhaps a pure-C implementation would be enough. IDK. I'm gonna sleep on it. G'night anon.
K so all we need to do is image search medium.com for an image with that header. If no results found (and the original filename isn't like medium's random naming bullshit) then we probably have a PK image.
Does stegdetect hit false positives?
Here's a medium article with the exact illumipepe image [positive ID by SHA] that's in your list.
https://medium. com/@Freequincy/right-wing-dove-squad-how-trash-dove-became-the-symbol-of-the-alt-right-c7794b84a48d
The pedo jewelry is the smallest image I can find that has the correct headers, gets a positive from stegdetect, and is not found on medium.com
Scratch that - wrong header
These bytes don't matter?
Thanks anon. Just wanted to confirm I understood that code correctly.
ROFL Holeee Sheit
https://archive.fo/VdhX8#selection-2949.0-2955.73
https://anonfile.com/EaG3B8fbb8/PKunknot.zip
This is what I'm using anon - single thread per instance though. You'll have to manually split your wordlists. It will automatically generate every permutation for a given charset and exit if a correct solution is found.
Run by calling the following on your command line:
java -cp bcprov-jdk15on-160.jar; q.Main %IMGNAME% %CHARSETFILE% %STARTINGWORD%