moyix

5 ms·

1.
by moyix 18d ago
Parts of this reminded me pleasantly of one of my favorite John Regehr posts, "Operant Conditioning by Software Bugs" https://blog.regehr.org/archives/861
2.
by moyix 1mo ago
They actually removed the temperature parameter starting with GPT-5.
3.
by moyix 4mo ago
I think you're confusing CVEs and vulnerabilities here? Mozilla (per their longstanding practice) grouped multiple vulnerabilities found internally under a small number of CVEs.
4.
Reverse Engineering SimTower(phulin.me)
4 points by moyix 5mo ago|1 comments
5.
by moyix 5mo ago
On hardened targets and Firecracker specifically, here's a recent vulnerability found by "Anthropic": https://aws.amazon.com/security/security-bulletins/2026-015-... Unfortunately it's unclear
6.
by moyix 6mo ago
It's limiting from the PoV of a developer who wants to ensure that their own code is free of all security issues. It is not limiting from the point of view of an attacker who just needs one good memory safety vuln to win.
7.
by moyix 6mo ago
This is true for a lot of things but for low-level code you can always fall back to "the intention is to not violate memory safety".
8.
by moyix 8mo ago
Also, unlike OpenAI, Anthropic's prompt caching is explicit (you set up to 4 cache "breakpoints"), meaning if you don't implement caching then you don't benefit from it.
9.
by moyix 8mo ago
There is filtering mentioned, it's just not done by a human: > I have written up the verification process I used for the experiments here, but the summary is: an exploit tends to involve building a capability to allow you to do some
10.
by moyix 1y ago
Note that MuZero did better than AlphaGo, without access to preprogrammed rules: https://en.wikipedia.org/wiki/MuZero
11.
by moyix 1y ago
There's also a FIDO standard in the works for how to export passkeys: https://blog.1password.com/fido-alliance-import-export-passk...
12.
by moyix 1y ago
The main difference is that all of the vulnerabilities reported here are real, many quite critical (XXE, RCE, SQLi, etc.). To be fair there were definitely a lot of XSS, but the main reason for that is that it's a really common vulnera
13.
by moyix 1y ago
All of these reports came with executable proof of the vulnerabilities – otherwise, as you say, you get flooded with hallucinated junk like the poor curl dev. This is one of the things that makes offensive security an actually good use case
14.
by moyix 1y ago
Wait a sec, I thought they were optional? > White Paper/Slide Deck/Supporting Materials (optional) > • If you have a completed white paper or draft, slide deck, or other supporting materials, you can optionally provide a lin
15.
by moyix 1y ago
Yeah, it's been very strange being on the other side of that after 10 years in academia! But it's totally reasonable for people to be skeptical when there's a bunch of money sloshing around. I'll see if I can get time to
16.
by moyix 1y ago
This is discussed in the post – many came down to individual programs' policies e.g. not accepting the vulnerability if it was in a 3rd party product they used (but still hosted by them), duplicates (another researcher reported the sam
17.
by moyix 1y ago
We've got a bunch of agent traces on the front page of the web site right now. We also have done writeups on individual vulnerabilities found by the system, mostly in open source right now (we did some fun scans of OSS projects found o
18.
by moyix 1y ago
You should come to my upcoming BlackHat talk on how we did this while avoiding false positives :D https://www.blackhat.com/us-25/briefings/schedule/#ai-agents...
19.
by moyix 1y ago
I made a CTF challenge based on that lovely feature of select() :D You could use the out-of-bounds bitset memory corruption to flip bits in an RSA public key in a way that made it factorable, generate the corresponding private key, and use
20.
by moyix 1y ago
With security vulnerabilities, you don't give the agent the ability to modify the potentially vulnerable software, naturally. Instead you make them do what an attacker would have to do: come up with an input that, when sent to the unmo
21.
by moyix 1y ago
He did do exactly what you say – except right after that, while reviewing the outputs, he found that it had also discovered a different 0day.
22.
by moyix 2y ago
One thing that is interesting is that this was anticipated by the OpenAI Codex paper (which led to GitHub Copilot) all the way back in 2021: > Users might be more inclined to accept the Codex answer under the assumption that the package
23.
by moyix 2y ago
I think the usual name is "overlay". At least, that's what Tim Gowers called the one he started :) https://gowers.wordpress.com/2015/09/10/discrete-analysis-an...
24.
by moyix 2y ago
I'm a bit confused, or maybe I've been doing it wrong. DEI-related things don't usually go in Broader Impacts, do they? When I've written grants, Broader Impacts was just generally for "how is your research going to
25.
by moyix 2y ago
We've been using them to find novel vulnerabilities in open source web apps. The past 4 posts here have details: - Auth bypass/arbitrary file read in Scoold: https://xbow.com/blog/xbow-scoold-vuln/ - SSR
26.
Finding a Critical Vulnerability in Scoold with AI(xbow.com)
4 points by moyix 2y ago|1 comments
27.
by moyix 2y ago
[This is more of a reply to a deleted reply to you, but I don't want my efforts to go to waste] Spatial memory safety is a reasonably common term in the security / PL field. You can see examples of it being used at least as far ba
28.
by moyix 2y ago
Note that the vulnerable extension is only enabled in the sqlite shell: > However, the generate_series extension is only enabled by default in the shell binary and not the library itself, so the impact of the issue is limited. https:&#x
29.
by moyix 2y ago
nmap can't really tell the difference between an open or a firewalled UDP port. For this specific vuln you can send it a packet like: echo "0 3 http://myserver:PORT/printers/foo " | nc -u target 631 And i
30.
by moyix 2y ago
Long agent trajectories, especially with command outputs.

More ›