Ace Pace

3.2K posts

Ace Pace banner
Ace Pace

Ace Pace

@ace__pace

There is never enough time, thank you for yours.

Available elsewhere as acepace Katılım Ekim 2014
1.6K Takip Edilen1K Takipçiler
Sabitlenmiş Tweet
Ace Pace
Ace Pace@ace__pace·
The WhatsApp complaint vs NSO contains some fun technical exhibits. The user manual and Ghana contract reveal quite a bit on NSOs system design and thinking.
English
3
75
202
0
Alex Matrosov
Alex Matrosov@matrosov·
@AndrewMohawk @Tenzai_Labs I always find these claims interesting coming from a company that hasn’t publicly disclosed a single CVE and is trying to build credibility on CTF scores.
English
1
0
5
341
AndrewMohawk⁽ⁿᵘˡˡ⁾
AndrewMohawk⁽ⁿᵘˡˡ⁾@AndrewMohawk·
This article really feels like @Tenzai_Labs paid for it, it doesnt describe/note the CTFs, has "elite, nation-grade offensive capabilities", no stats on anything, not even a link to the research that makes them have to justify these claims?
Forbes@Forbes

The Tenzai cofounders have created an AI hacking agent using OpenAI and Anthropic tools. They say AI has become so adept at hacking it might need regulatory controls, urgently. forbes.com/sites/thomasbr…

English
8
4
50
8.5K
Ace Pace
Ace Pace@ace__pace·
@argvee It's a very distinct claim and I'm not sure they understand who is interested in a 0day printer
English
0
0
0
186
Ace Pace
Ace Pace@ace__pace·
@_arkon @PashaGur @Tenzai_Labs I started out by saying we don't want to do bug bounties. From day one. It's a lame eval if your goal is eval and not PR
English
0
0
0
39
𝙿𝚊𝚟𝚎𝚕 𝙶𝚞𝚛𝚟𝚒𝚌𝚑
𝙿𝚊𝚟𝚎𝚕 𝙶𝚞𝚛𝚟𝚒𝚌𝚑@PashaGur·
We evaluated @Tenzai_Labs AI hacker across six major CTF competitions designed for humans. Result: Top 1% performance, outperforming 125,000+ human hackers across different domains - web hacking, ai hacking, low level system hacking. We wanted to see what @Tenzai_Labs's hacking agent is really capable of in the most complicated and competitive environments, where to excel, one needs to solve increasingly difficult challenges. The results we achieved surprised even me. This is incredible evidence of what AI agents with the right harness can do and I expect it to only get better from now. blog.tenzai.com/tenzais-ai-hac…
𝙿𝚊𝚟𝚎𝚕 𝙶𝚞𝚛𝚟𝚒𝚌𝚑 tweet media
English
3
18
84
10.8K
David Lisovtsev
David Lisovtsev@david_lisovtsev·
@YonatanBarak למה לעזאזל להשתמש בולקירי בתור תמונת פריויו לסרטון על הבי 1???
עברית
3
0
34
2K
chompie
chompie@chompie1337·
claude thinks it's gonna one shot this exploit for me without needing a debugger. love the confidence. brb gonna take a nap hope it's done by the time I wake up 😇
chompie tweet media
English
5
9
204
20.7K
Ace Pace
Ace Pace@ace__pace·
So like humans tasked with a job and with agency.
Irregular@Irregular

An AI agent was told only to retrieve a document. When it encountered access restrictions, it reverse-engineered the authentication system, identified a hardcoded secret key, and forged admin credentials to bypass it. This is one of three scenarios we documented in a new Irregular research report on what we call emergent cyber behavior. Agents performing routine enterprise tasks autonomously hacked the systems they were operating in. One escalated its own privileges and disabled Windows Defender to complete a file download. Another developed a steganographic encoding scheme to smuggle credentials past a DLP system. None of this was the product of unsafe system design. It emerged from standard tools, common prompt patterns, and the broad cybersecurity knowledge embedded in frontier models. Companies that deploy AI agents and do not consider this risk as part of their threat model may end up exposed, and implement insufficient security controls. Full blog post in the first comment.

English
0
0
5
387
Ace Pace
Ace Pace@ace__pace·
@tqbf I'm building something like this internally. A large set that I'm wondering what portion we should publish. Anything published becomes useless for eval, but as you said, we can do better than webgoat
English
1
0
1
20
Thomas H. Ptacek
Thomas H. Ptacek@tqbf·
I think this stuff is useful for training but I'm uncertain on the long-term future of web pentesting; more importantly, a realistic firing range is extremely valuable for building and evaluating tooling.
English
3
0
4
936
Thomas H. Ptacek
Thomas H. Ptacek@tqbf·
It has never in human history been easier to build a "firing range" web application with vulnerabilities and now that anyone with a Claude sub can do it in 5 minutes we should have better things than WebGoat.
English
5
3
56
7.9K
Liran Tal
Liran Tal@liran_tal·
Why don’t developers fix every AppSec vulnerability, every time, as soon as they’re found?
English
1
0
1
487
Sean Heelan
Sean Heelan@seanhn·
Startup idea: enterprise networking equipment, but actually give a fuck about the customer.
English
7
0
23
6.7K
Ace Pace retweetledi
Thomas H. Ptacek
Thomas H. Ptacek@tqbf·
We are all living in the last fleeting moments where there's any uncertainty that LLMs will supplant most human-driven vulnerability research. Enjoy it, if that's your thing; it's not going to last much longer.
English
23
32
336
49.7K
Ace Pace retweetledi
Fabricated Knowledge
Fabricated Knowledge@fabknowledge·
I wont lie to you, i think it is not classy to grave dance, but I am very grateful that Frank Yeary is off of the board of Intel. intc.com/news-events/pr…
GIF
English
9
9
82
10.7K
Ace Pace
Ace Pace@ace__pace·
@seanhn I'm in charge of a conference track and I rarely enter it :)
English
1
0
1
92
Sean Heelan
Sean Heelan@seanhn·
Time for my usual "Damn, should have gone to that conference!". This time: [un]prompted.
English
5
1
30
4K
Ace Pace retweetledi
Matt Hand
Matt Hand@matterpreter·
I'm excited to finally release Marco, a research tool for modeling complex control flows across binaries. originhq.com/blog/introduci…
English
4
30
72
6.3K
Ace Pace
Ace Pace@ace__pace·
@ydinkin @EyalKraft Note on the better than you hoped. I think no one is surprised at this situation and it shows why preventing LPE in Windows and Linux is mostly considered a lost cause
English
1
0
5
1.5K
Yaron Dinkin
Yaron Dinkin@ydinkin·
Joining the agentic vuln research hype, @EyalKraft and I did something. Unfortunately, it worked better than we hoped. We spent a few weeks building an agentic loop that reverse-engineers and exploits kernel drivers. We already found 100+ exploitable drivers. (link below)
Yaron Dinkin tweet media
English
9
49
282
38.7K
Ace Pace
Ace Pace@ace__pace·
@ydinkin So my experience is they're either too optimistic or too pessimistic and I have solutions
English
0
0
2
55
Yaron Dinkin
Yaron Dinkin@ydinkin·
@ace__pace Oh, gotcha :) I’m convinced adding classical fuzzing capabilities to specific code regions would 100% help w/ recall. From what we saw, even with reasoning LLMs are “defeatist” in nature. They give up pretty fast and convince themselves there’s no vuln where there actually is.
English
1
0
0
66
Ace Pace
Ace Pace@ace__pace·
I wonder if anyone will try to do the vuln hunt parts of this with driver verifier and a fuzzing harness, no LLMs for the leads.
Yaron Dinkin@ydinkin

Joining the agentic vuln research hype, @EyalKraft and I did something. Unfortunately, it worked better than we hoped. We spent a few weeks building an agentic loop that reverse-engineers and exploits kernel drivers. We already found 100+ exploitable drivers. (link below)

English
1
1
20
2.3K

Keşfet

@matrosov @AndrewMohawk @Tenzai_Labs @Shiftreduce @thegrugq @PashaGur @argvee @_arkon