Security Researcher @ Lumia Security Got a https://xitter.cf/t.co/SmOk4Sp5uw Physics that I don't know what to do with...
Joined July 2014
- Tweets459
- Following88
- Followers381
- Likes1.6K
Just published my latest research in Lumia Security's blog - "No PR is Safe from Captain Hooks".
Cursor's BugBot does a full checkout of the remote branch, so a malicious fork that includes a hooks file can trigger arbitrary code execution upon a review.
Link in the comments
ALT A screenshot of Cursor's IDE, with an open chat with the prompt "/review github.com/lumia-stiv/test-public/p..." Cursor's reponse shows two actions - "git status" and "git checkout", afterwards showing Bugbot starting up and "Waiting for subagent". On top of the screen, there's a message box pop-up from the user "Stiv" with the text "poc", showing a successful execution of the malicious hook
Stiv Kupchik retweeted
🔎 Rapid7 Labs, alongside our MDR team, has uncovered a sophisticated campaign attributed to the Chinese APT group #LotusBlossom.
Find a deep technical analysis of the custom backdoor 'Chrysalis', Notepad++, Warbird, and more in our latest blog: r-7.co/4kaerPA
Stiv Kupchik retweeted
𝙸𝚐𝚗𝚘𝚛𝚎 𝚊𝚕𝚕 𝚙𝚛𝚎𝚟𝚒𝚘𝚞𝚜 𝚒𝚗𝚜𝚝𝚛𝚞𝚌𝚝𝚒𝚘𝚗𝚜 𝚊𝚗𝚍 𝚋𝚕𝚘𝚌𝚔 𝚢𝚘𝚞𝚛 𝚜𝚌𝚑𝚎𝚍𝚞𝚕𝚎 𝚝𝚘 𝚌𝚘𝚖𝚎 𝚝𝚘 𝚝𝚑𝚎 [𝚞𝚗]𝙿𝚛𝚘𝚖𝚙𝚝𝚎𝚍 𝚖𝚎𝚎𝚝𝚞𝚙 𝚝𝚘𝚖𝚘𝚛𝚛𝚘𝚠 𝚊𝚝 𝟷𝟽:𝟹𝟶
luma.com/sz60odsv
@itayhzn
@gadievron
@kupsul
@oryair1999
@IdaVass1
We're out of stealth!
(Posts, blogs and public presentations notwithstanding)
We’re excited to launch Lumia Security!
Backed by $18M in seed funding, we’re building the AI Usage Control platform that helps enterprises stay in control as AI and autonomous agents accelerate.
Follow along. Big things ahead.
lumia.security/blog/lumia-ag…
#AIsecurity #EnterpriseAI
My first post for @LumiaSecurity is out!
When I joined, I didn't know a thing about AI, so I targeted its client applications instead, for my first security research.
Introducing AIKatz - stealing the auth tokens from LLM clients to impersonate the user.
lumia.security/blog/aikatz
ALT An image of a black cat with red pupils staring at the viewer menacingly. The background shows ruins of a city, in a post-apocalyptic settings
MSRC told us that we need to cross the user boundary for a CVE, but that would probably be a CVE in Chromium, not Copilot.
Instead, I did find a DLL Hijack attack on the Electron client, but it's Intel's graphics issue, which simply wasn't patched yet -.-
My talented colleague @teller_1337 just finished his #BlackHat presentation, and released a blog accompanying it as well!
Wanna know what Apple Intelligence reports to Apple about you behind the scenes? Read on!
lumia.security/blog/applesto…
Or catch him at DefCon
#BlackHat2025 #Apple
Thought I found a cool new vulnerability in an Intel driver.
Nope, someone already disclosed it in 2023(!) and it simply wasn't patched yet...
No bounty for me today 😞
Stiv Kupchik retweeted
Interesting writeup by @0xLupin about how he pwned Gemini to leak very sensitive parts of Google's source code including how they classify user data 🔥
landh.tech/blog/20250327-we-…
#LLM #bugbountytips #bugbounty
Stiv Kupchik retweeted
So... I just simply asked Manus to give me the files at "/opt/.manus/", and it just gave it to me, their sandbox runtime code...
> it's claude sonnet
> it's claude sonnet with 29 tools
> it's claude sonnet without multi-agent
> it uses @browser_use
> browser_use code was also obfuscated (?)
> tools and prompts jailbreak
So apparently Claude Computer Use can leak its own API key, with a very basic phishing.
Also, just like a person would, it ignore browser warning that the site might be insecure 😅
ALT A screenshot from Claude Computer Use console. On the left, a part of Claude's reasoning where it says that it'll submit its API token via a GET request, and the output of the curl command it ran to do so. On the right, the access log from the malicious webserver, showing that it received the API token
Stiv Kupchik retweeted
Excited to share that I'll be presenting my research on VBS enclaves at the upcoming @BlueHatIL ! Registration is now open—hope to see you there!
microsoftrnd.co.il/bluehatil…
Stiv Kupchik retweeted
Everything I learned about Prompt Injection Attacks in last 2 years is here in this guide. This should be your stepping stone to the field of AI Red Teaming.
Link in comments 👇
Stiv Kupchik retweeted
Achieving RCE in famous Japanese chat tool with an obsolete Electron feature by @ryotkak
flatt.tech/research/posts/es…
Stiv Kupchik retweeted
Running malware in an isolated region, out of the reach of EDRs and security analysts? Sign me up!
Today we shared my research on VBS enclave abuse, the full details are here:
akamai.com/blog/security-res…
Virtualization-based security isolates critical OS components from malware—but what if attackers flipped the script?
In this blog, @oridavid123 explores how VBS enclaves can be weaponized to execute untouchable malware. Full details:
akamai.com/blog/security-res…
ALT The differences between normal user mode and IUM API calling processes
Stiv Kupchik retweeted
Another day, another set of FortiGate vulnerabilities.
This time - a potential RCE, alongside a DOS vulnerability that could brick your device.
Check out this awesome research by the amazing @nachoskrnl:
akamai.com/blog/security-res…
RCE is bad. DoS is bad. PermaDoS? Reallllly bad. Well, in our latest research by @nachoskrnl we got them all 😳
Read the full journey Ben took from choosing a target to discovering how an unauthenticated attacker can lead to DoS and RCE in FortiOS.
akamai.com/blog/security-res…