The ai harness I open sourced a few days ago has now an update, now it also runs foundry tests on the hypothesis it built, then we can confirm a finding.
Working on the go as I hit some issues myself during the hunts and audits, trying to solve by building what I wish I had.
@pashov @asen_sec
github.com/SamarthShukla17/h…
Hound is putting in the work for me lately.
It has been a really good tool till now and I am looking forward to calibrate it for specific hunts and audits, that would come pretty handy in contests and audits.
Till now have lead me to around 10+ direct/indirect finds, most of them came out dupes yet they were never publicly known or disclosed till that point so its a pretty great result.
github.com/SamarthShukla17/h…
Just finished another level!
I am hacking Dex in the SCH Exploit Challenge.
Can you beat me? 😉 smartcontractshacking.com/to…
Security Researchers obsessing over tokenmaxxing and modelmaxxing are not going to make it
Been seeing whitehats go beastmode on a $200 gpt/opus plan, crushing findings left & right - get your harness right, the model doesn't matter and more tokens is rarely the answer🙏
I have open sourced the ai harness that I made for security research, it produces the correct areas to look for the vulnerabilities and fails in spotting the correct bugs directly, still in early phase.
Thanks to @asen_sec for the guidance on custom tooling via his blogs
github.com/SamarthShukla17/h…
Here's what my hacker dashboard looks like.
@HackenProof
hackenproof.com/hackers/Sama…









