Show HN: Strix - Open-source AI hackers for your apps

by ahmedallam2 • 11 months ago

102 15 11 months ago

waihtis2• 11 months ago

The joke is that Xbow only works because they have close to 100 employees operating the software

_pdp_• 11 months ago

You are joking, but there was actually a very popular enterprise SAST tool that used to offer a "cloud" version of their software. It worked by having someone from their team manually download the zip file of your code, run it through their desktop software, and then upload the results back to make them visible in the web portal.

ericmcer• 11 months ago

That's a totally valid and useful way to validate an idea. After a few months of manual labor they will have a good idea of how/what to build and if it is even worth building.

tptacek3• 11 months ago

I know who you're talking about, but also: this is the joke about basically every hosted SAST and DAST tool. I call it the "Oompa Loompa" model of security products.

guhcampos• 11 months ago

"XBOW is an AI-powered penetration testing platform that delivers human-level security testing at machine speed."

At least they're not lying right? It's just people using computers.

waihtis2• 11 months ago

yeah i think XBOW is fairly transparent about it, doesn't stop the online influencers from claiming "an AI is now the #1 hacker on Hackerone"

armanj• 11 months ago

Took a while to notice it's xbow and not xbox

tptacek3• 11 months ago

This is a neat project, I don't know why you'd want to set it up with this comparison to Xbow. As someone who works (worked? I'm non-ironically still trying to figure out if I belong in this space post-AI!) in this space and knows some of the actors, I'm pretty sure there's more to Xbow than ~1800 lines of prompts. Like: this is your RCE template prompt:

https://github.com/usestrix/strix/blob/main/strix/prompts/vu...

... and this is great, I'm not dunking, but pretty basic?

We just had the DARPA AIxCC results come in, and those systems are (1) open source and (2) presumably simpler/less polished than Xbow (some of the authors will be quick to tell you that they're doing PoC work, not product development), and (3) they're more complicated than this.

Again, to be super clear: I think there's a huge amount of potential in building something like this up. Nessus was much simpler than ISS when it first shipped, but you'd rather be Nessus than an ISS scanner developer! I'm just: why set this bar for your project?

Best of luck with this!

captn3m0• 11 months ago

Buttercup, from the Trail of Bits team has a nice writeup + FOSS : https://blog.trailofbits.com/2025/08/08/buttercup-is-now-ope...

thegeomaster• 11 months ago

Seems heavily vibe coded, down to the Claude-generated README and a lot of the LLM prompts themselves (which I have found works very poorly compared to human-written prompts). While none of this is necessarily bad, it requires a higher burden of proof that it actually works beyond toy problems [0]. I think everyone would appreciate some examples of vulnerabilities it can find. The missing JWT check showcased in the screenshot would've probably been caught with ordinary AI code review, so to my eye that by itself is not persuasive.

Good luck!

[0]: Why I say this --- a 10kLOC piece of software that was mostly human-written would require a large amount of testing, even manual, to ensure that it works, reliably, at all. All this testing and experimentation would naturally force a certain depth of exploration for the approach, the LLM prompts, etc across a variety of usecases. A mostly AI-written codebase of this size would've required much less testing to get it to "doesn't crash and runs reliably", and so this depth is not a given anymore.

Show HN: Strix - Open-source AI hackers for your apps

Related Stories