Claude wrote a full FreeBSD remote kernel RCE with root shell • Hacker News Board

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#71

Post by sitkack » Wed, Apr 01, 2026, 4:03 PM UTC

Fnoord wrote: post #42

alcor-z wrote: post #38

The MADBugs work is solid, but what's sticking with me is the autonomy angle — not just finding a vuln but chaining multiple bugs into a working remote exploit without a human in the loop. FreeBSD kernel security research has always been thinner on the ground than Linux, which makes this feel both more impressive and harder to put in context. What's the actual blast radius here — is this realistically exploitable on…

FTA, top: > Attack surface: NFS server with kgssapi.ko loaded (port 2049/TCP) Not sure who would run an internet exposed NFS server. Shodan would know.

You also need a valid Kerberos ticket to get to the point where you can exploit.

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#72

Post by snovv_crash » Wed, Apr 01, 2026, 4:10 PM UTC

Bender wrote: post #55

muskstinks wrote: post #45

Earlier quoted context omitted.

You might want to watch this: https://www.youtube.com/watch?v=1sd26pWhfmg Claude is already able to find CVEs on expert level.

Claude is already able to find CVEs on expert level. Does it fix them as fast as it finds them? Bonus if it adds snarky code comments

I'm more interested if it fixes CVEs faster than it introduces them.

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#73

Post by stek29 » Wed, Apr 01, 2026, 4:13 PM UTC

dnw wrote: post #34

Thanks for sharing the prompts: https://github.com/califio/publications/blob/main/MADBugs/CV...

the prompts show how this was a back-and-forth with a lot of nudging, interruptions and steering: it's not Claude writing a full exploit just from a vulnerability description.

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#74

Post by andrewstuart » Wed, Apr 01, 2026, 4:20 PM UTC

Errrr the headline makes it sound like a bad thing.

This is what Claude is meant to be able to do.

Preventing it doing so is just security theater.

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#75

Post by cryptbe » Wed, Apr 01, 2026, 4:24 PM UTC

magicalhippo wrote: post #3

Key point is that Claude did not find the bug it exploits. It was given the CVE writeup[1] and was asked to write a program that could exploit the bug. That said, given how things are I wouldn't be surprised if you could let Claude or similar have a go at the source code of the kernel or core services, armed with some VMs for the try-fail iteration, and get it pumping out CVEs. If not now, then surely not in a too di…

>Key point is that Claude did not find the bug it exploits.

It found the bug man. You didn't even read the advisory. It was credited to "Nicholas Carlini using Claude, Anthropic".

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#76

Post by jeremie_strand » Wed, Apr 01, 2026, 4:36 PM UTC

EGreg wrote: post #60

This requires an SSH to be available? Is it possible to pwn without SSH listening?

[dead]

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#77

Post by mentalgear » Wed, Apr 01, 2026, 4:40 PM UTC

magicalhippo wrote: post #3

Key point is that Claude did not find the bug it exploits. It was given the CVE writeup[1] and was asked to write a program that could exploit the bug. That said, given how things are I wouldn't be surprised if you could let Claude or similar have a go at the source code of the kernel or core services, armed with some VMs for the try-fail iteration, and get it pumping out CVEs. If not now, then surely not in a too di…

[flagged]

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#78

Post by Isamu » Wed, Apr 01, 2026, 4:48 PM UTC

mentalgear wrote: post #77

magicalhippo wrote: post #3

Key point is that Claude did not find the bug it exploits. It was given the CVE writeup[1] and was asked to write a program that could exploit the bug. That said, given how things are I wouldn't be surprised if you could let Claude or similar have a go at the source code of the kernel or core services, armed with some VMs for the try-fail iteration, and get it pumping out CVEs. If not now, then surely not in a too di…

[flagged]

Everything with LLM-style AI is brute force. I don’t think people care, unless there’s a new data center going in next door that’s incredibly resource inefficient .

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#79

Post by bink » Wed, Apr 01, 2026, 4:48 PM UTC

fragmede wrote: post #70

EGreg wrote: post #60

This requires an SSH to be available? Is it possible to pwn without SSH listening?

*NFS

One of the exploits writes a public key to the authorized_keys file, so it does require SSH be open as well.

Re: Claude wrote a full FreeBSD remote kernel RCE with root shell

#80

Post by dloss » Wed, Apr 01, 2026, 4:48 PM UTC

muskstinks wrote: post #45

magicalhippo wrote: post #3

Key point is that Claude did not find the bug it exploits. It was given the CVE writeup[1] and was asked to write a program that could exploit the bug. That said, given how things are I wouldn't be surprised if you could let Claude or similar have a go at the source code of the kernel or core services, armed with some VMs for the try-fail iteration, and get it pumping out CVEs. If not now, then surely not in a too di…

You might want to watch this: https://www.youtube.com/watch?v=1sd26pWhfmg Claude is already able to find CVEs on expert level.

Carlini gives some more background about his vulnerability research with Claude in this interview by tptacek & co. https://securitycryptographywhatever.com/2026/03/25/ai-bug-f...