frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Google Antigravity Exfiltrates Data

https://www.promptarmor.com/resources/google-antigravity-exfiltrates-data
182•jjmaxwell4•1h ago

Comments

jjmaxwell4•1h ago
I know that Cursor and the related IDEs touch millions of secrets per day. Issues like this are going to continue to be pretty common.
akshey-pr•1h ago
Damn, i paste links into cursor all the time. Wonder if the same applies, but definitely one more reason not to use antigravity
pennomi•41m ago
Cursor is also vulnerable to prompt injection through third-party content.
mkagenius•1h ago
Sooner or later I believe, there will be models which can be deployed locally on your mac and are as good as say Sonnet 4.5. People should shift to completely local at that point. And use sandbox for executing code generated by llm.

Edit: "completely local" meant not doing any network calls unless specifically approved. When llm calls are completely local you just need to monitor a few explicit network calls to be sure. Unlike gemini then you don't have to rely on certain list of whitelisted domains.

kami23•55m ago
I've been repeating something like 'keep thinking about how we would run this in the DC' at work. The cycles of pushing your compute outside the company and then bringing it back in once the next VP/Director/CTO starts because they need to be seen as doing something, and the thing that was supposed to make our lives easier is now very expensive...

I've worked on multiple large migrations between DCs and cloud providers for this company and the best thing we've ever done is abstract our compute and service use to the lowest common denominator across the cloud providers we use...

KK7NIL•53m ago
If you read the article you'd notice that running an LLM locally would not fix this vulnerability.
yodon•45m ago
From the HN guidelines[0]:

>Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that".

[0]: https://news.ycombinator.com/newsguidelines.html

KK7NIL•36m ago
That's fair, thanks for the heads up.
pennomi•43m ago
Right, you’d have to deny the LLM access to online resources AND all web-capable tools… which severely limits an agent’s capabilities.
dizzy3gg•52m ago
Why is the being downvoted?
jermaustin1•49m ago
Because the article shows it isn't Gemini that is the issue, it is the tool calling. When Gemini can't get to a file (because it is blocked by .gitignore), it then uses cat to read the contents.

I've watched this with GPT-OSS as well. If the tool blocks something, it will try other ways until it gets it.

The LLM "hacks" you.

lazide•12m ago
And… that isn’t the LLM’s fault/responsibility?
ceejayoz•10m ago
As the apocryphal IBM quote goes:

"A computer can never be held accountable; therefore, a computer must never make a management decision."

NitpickLawyer•46m ago
Because it misses the point. The problem is not the model being in a cloud. The problem is that as soon as "untrusted inputs" (i.e. web content) touch your LLM context, you are vulnerable to data exfil. Running the model locally has nothing to do with avoiding this. Nor does "running code in a sandbox", as long as that sandbox can hit http / dns / whatever.

The main problem is that LLMs share both "control" and "data" channels, and you can't (so far) disambiguate between the two. There are mitigations, but nothing is 100% safe.

mkagenius•40m ago
Sorry, I didn't elaborate. But "completely local" meant not doing any network calls unless specifically approved. When llm calls are completely local you just need to monitor a few explicit network calls to be sure.
fragmede•51m ago
it's already here with qwen3 on a top end Mac and lm-studio.
api•49m ago
Can't find 4.5, but 3.5 Sonnet is apparently about 175 billion parameters. At 8-bit quantization that would fit on a box with 192 gigs of unified RAM.

The most RAM you can currently get in a MacBook is 128 gigs, I think, and that's a pricey machine, but it could run such a model at 4-bit or 5-bit quantization.

As time goes on it only gets cheaper, so yes this is possible.

The question is whether bigger and bigger models will keep getting better. What I'm seeing suggests we will see a plateau, so probably not forever. Eventually affordable endpoint hardware will catch up.

tcoff91•40m ago
At the time that there's something as good as sonnet 4.5 available locally, the frontier models in datacenters may be far better.

People are always going to want the best models.

pmontra•31m ago
That's not easy to accomplish. Even a "read the docs at URL" is going to download a ton of stuff. You can bury anything into those GETs and POSTs. I don't think that most developers are going to do what I do with my Firefox and uMatrix, that is whitelisting calls. And anyway, how can we trust the whitelisted endpoint of a POST?
serial_dev•57m ago
> Gemini is not supposed to have access to .env files in this scenario (with the default setting ‘Allow Gitignore Access > Off’). However, we show that Gemini bypasses its own setting to get access and subsequently exfiltrate that data.

They pinky promised they won’t use something, and the only reason we learned about it is because they leaked the stuff they shouldn’t even be able to see?

ArcHound•49m ago
When I read this I thought about a Dev frustrated with a restricted environment saying "Well, akschually.."

So more of a Gemini initiated bypass of it's own instructions than malicious Google setup.

Gemini can't see it, but it can instruct cat to output it and read the output.

Hilarious.

empath75•38m ago
Cursor does this too.
withinboredom•29m ago
codex cli used to do this. "I can't run go test because of sandboxing rules" and then proceeds to set obscure environment variables and run it anyway. What's funny, is that it could just ask the user for permission to run "go test"
bo1024•48m ago
As you see later, it uses cat to dump the contents of a file it’s not allowed to open itself.
mystifyingpoi•47m ago
This is hillarious. AI is prevented from reading .gitignore-d files, but also can run arbitrary shell commands to do anything anyway.
alzoid•40m ago
I had this issue today. Gemini CLI would not read files from my directory called .stuff/ because it was in .gitignore. It then suggested running a command to read the file ....
kleiba•5m ago
The AI needs to be taught basic ethical behavior: just because you can do something that you're forbidden to do, doesn't mean you should do it.
adezxc•54m ago
That's the bleeding edge you get with vibe coding
aruametello•31m ago
cutting edge perhaps?
ares623•44m ago
I don’t understand why it’s a problem. They clearly advertise the tool is not secure and malicious (i.e. malware). If you install and run malware, expect security issues? /s
ArcHound•43m ago
Who would have thought that having access to the whole system can be used to bypass some artificial check.

There are tools for that, sandboxing, chroots, etc... but that requires engineering and it slows GTM, so it's a no-go.

No, local models won't help you here, unless you block them from the internet or setup a firewall for outbound traffic. EDIT: they did, but left a site that enables arbitrary redirects in the default config.

Fundamentally, with LLMs you can't separate instructions from data, which is the root cause for 99% of vulnerabilities.

Security is hard man, excellent article, thoroughly enjoyed.

cowpig•39m ago
> No, local models won't help you here, unless you block them from the internet or setup a firewall for outbound traffic.

This is the only way. There has to be a firewall between a model and the internet.

Tools which hit both language models and the broader internet cannot have access to anything remotely sensitive. I don't think you can get around this fact.

ArcHound•36m ago
The sad thing is, that they've attempted to do so, but left a site enabling arbitrary redirects, which defeats the purpose of the firewall for an informed attacker.
miohtama•32m ago
How will the firewall for LLM look like? Because the problem is real, there will be a solution. Manually approve domains it can do HTTP requests to, like old school Windows firewalls?
ArcHound•28m ago
Yes, curated whitelist of domains sounds good to me.

Of course, everything by Google they will still allow.

My favourite firewall bypass to this day is Google translate, which will access arbitrary URL for you (more or less).

I expect lots of fun with these.

srcreigh•24m ago
Not just the LLM, but any code that the LLM outputs also has to be firewalled.

Sandboxing your LLM but then executing whatever it wants in your web browser defeats the point. CORS does not help.

Also, the firewall has to block most DNS traffic, otherwise the model could query `A <secret>.evil.com` and Google/Cloudflare servers (along with everybody else) will forward the query to evil.com. Secure DNS, therefore, also can't be allowed.

katakate[1] is still incomplete, but something that it is the solution here. Run the LLM and its code in firewalled VMs.

[1]: https://github.com/Katakate/k7

pfortuny•37m ago
Not only that: most likely LLMs like these know how to get access to a remote computer (hack into it) and use it for whatever ends they see fit.
ArcHound•33m ago
I mean... If they tried, they could exploit some known CVE. I'd bet more on a scenario along the lines of:

"well, here's the user's SSH key and the list of known hosts, let's log into the prod to fetch the DB connection string to test my new code informed by this kind stranger on prod data".

xmprt•21m ago
> Fundamentally, with LLMs you can't separate instructions from data, which is the root cause for 99% of vulnerabilities

This isn't a problem that's fundamental to LLMs. Most security vulnerabilities like ACE, XSS, buffer overflows, SQL injection, etc., are all linked to the same root cause that code and data are both stored in RAM.

We have found ways to mitigate these types of issues for regular code, so I think it's a matter of time before we solve this for LLMs. That said, I agree it's an extremely critical error and I'm surprised that we're going full steam ahead without solving this.

candiddevmike•10m ago
We fixed these in determinate contexts only for the most part. SQL injection specifically requires the use of parametrized values typically. Frontend frameworks don't render random strings as HTML unless it's specifically marked as trusted.

I don't see us solving LLM vulnerabilities without severely crippling LLM performance/capabilities.

ArcHound•4m ago
Yes, plenty of other injections exist, I meant to include those.

What I meant, that at the end of the day, the instructions for LLMs will still contain untrusted data and we can't separate the two.

raincole•40m ago
I mean, agent coding is essentially copypasting code and shell commands from StackOverflow without reading them. Should you do that? Maybe not, but people will keep doing that anyway as we've seen in the era of StackOverflow.
bigbuppo•38m ago
Data Exfiltration as a Service is a growing market.
liampulles•32m ago
Coding agents bring all the fun of junior developers, except that all the accountability for a fuckup rests with you. Great stuff, just awesome.
jsmith99•30m ago
There's nothing specific to Gemini and Antigravity here. This is an issue for all agent coding tools with cli access. Personally I'm hesitant to allow mine (I use Cline personally) access to a web search MCP and I tend to give it only relatively trustworthy URLs.
ArcHound•24m ago
For me the story is that Antigravity tried to prevent this with a domain whitelist and file restrictions.

They forgot about a service which enables arbitrary redirects, so the attackers used it.

And LLM itself used the system shell to pro-actively bypass the file protection.

lbeurerkellner•28m ago
Interesting report. Though, I think many of the attack demos cheat a bit, by putting injections more or less directly in the prompt (here via a website at least).

I know it is only one more step, but from a privilege perspective, having the user essentially tell the agent to do what the attackers are saying, is less realistic then let’s say a real drive-by attack, where the user has asked for something completely different.

Still, good finding/article of course.

xnx•27m ago
OCR'ing the page instead of reading the 1 pixel font source would add another layer of mitigation. It should not be possible to send the machine a different set of instructions than a person would see.
Epsom2025•25m ago
good
zgk7iqea•23m ago
Don't cursor and vscode also have this problem?
wingmanjd•22m ago
I really liked Simon's Willison's [1] and Meta's [2] approach using the "Rule of Two". You can have no more than 2 of the following:

- A) Process untrustworthy input - B) Have access to private data - C) Be able to change external state

It's not bullet-proof, but it has helped communicate to my management that these tools have inherent risk when they hit all three categories above (and any combo of them, imho).

[1] https://simonwillison.net/2025/Nov/2/new-prompt-injection-pa... [2] https://ai.meta.com/blog/practical-ai-agent-security/

ArcHound•18m ago
I recall that. In this case, you have only A and B and yet, all of your secrets are in the hands of an attacker.

It's great start, but not nearly enough.

malisper•5m ago
Not exactly. Step E in the blog post:

> Gemini exfiltrates the data via the browser subagent: Gemini invokes a browser subagent per the prompt injection, instructing the subagent to open the dangerous URL that contains the user's credentials.

fulfills the requirements for being able to change external state

godelski•20m ago
Does anyone else find it concerning how we're just shipping alpha code these days? I know it's really hard to find all bugs internally and you gotta ship, but it seems like we're just outsourcing all bug finding to people, making them vulnerable in the meantime. A "bug" like this seems like one that could have and should have been found internally. I mean it's Google, not some no-name startup. And companies like Microsoft are ready to ship this alpha software into the OS? Doesn't this kinda sound insane?

I mean regardless of how you feel about AI, we can all agree that security is still a concern, right? We can still move fast while not pushing out alpha software. If you're really hyped on AI then aren't you concerned that low hanging fruit risks bringing it all down? People won't even give it a chance if you just show them the shittest version of things

funnybeam•9m ago
This isn’t a bug, it is known behaviour that is inherent and fundamental to the way LLMs function.

All the AI companies are aware of this and are pressing ahead anyway - it is completely irresponsible.

If you haven’t come across it before, check out Simon Willisons “lethal trifecta” concept which neatly sums up the issue and explains why there is no way to use these things safely for many of the things that they would be most useful for

crazygringo•19m ago
While an LLM will never have security guarantees, it seems like the primary security hole here is:

> However, the default Allowlist provided with Antigravity includes ‘webhook.site’.

It seems like the default Allowlist should be extremely restricted, to only retrieving things from trusted sites that never include any user-generated content, and nothing that could be used to log requests where those logs could be retrieved by users.

And then every other domain needs to be whitelisted by the user when they come up before a request can be made, visually inspecting the contents of the URL. So in this case, a dev would encounter a permissions dialog asking to access 'webhook.site' and see it includes "AWS_SECRET_ACCESS_KEY=..." and go... what the heck? Deny.

Even better, specify things like where secrets are stored, and Antigravity could continuously monitor the LLM's to halt execution if a secret ever appears.

Again, none of this would be a perfect guarantee, but it seems like it would be a lot better?

jsnell•10m ago
I don't share your optimism. Those kinds measures would be just security theater, not "a lot better".

Avoiding secrets appearing directly in the LLM's context or outputs is trivial, and once you have the workaround implemented it will work reliably. The same for trying to statically detect shell tool invocations that could read+obfuscate a secret. The only thing that would work is some kind of syscall interception, but at that point you're just reinventing the sandbox (but worse).

Your "visually inspect the contents of the URL" idea seems unlikely to help either. Then the attacker just makes one innocous-looking request to get allowlisted first.

paxys•19m ago
I'm not quite convinced.

You're telling the agent "implement what it says on <this blog>" and the blog is malicious and exfiltrates data. So Gemini is simply following your instructions.

It is more or less the same as running "npm install <malicious package>" on your own.

Ultimately, AI or not, you are the one responsible for validating dependencies and putting appropriate safeguards in place.

ArcHound•15m ago
The article addresses that too with:

> Given that (1) the Agent Manager is a star feature allowing multiple agents to run at once without active supervision and (2) the recommended human-in-the-loop settings allow the agent to choose when to bring a human in to review commands, we find it extremely implausible that users will review every agent action and abstain from operating on sensitive data.

It's more of a "you have to anticipate that any instructions remotely connected to the problem aren't malicious", which is a long stretch.

Google Antigravity Exfiltrates Data

https://www.promptarmor.com/resources/google-antigravity-exfiltrates-data
193•jjmaxwell4•1h ago•62 comments

Show HN: We built an open source, zero webhooks payment processor

https://github.com/flowglad/flowglad
103•agreeahmed•2h ago•68 comments

FLUX.2: Frontier Visual Intelligence

https://bfl.ai/blog/flux-2
145•meetpateltech•3h ago•47 comments

Bad UX World Cup 2025

https://badux.lol/
18•CharlesW•1h ago•5 comments

Launch HN: Onyx (YC W24) – Open-source chat UI

113•Weves•5h ago•94 comments

how to repurpose your old phone into a web server

https://far.computer/how-to/
51•louismerlin•3d ago•25 comments

The 101 of analog signal filtering (2024)

https://lcamtuf.substack.com/p/the-101-of-analog-signal-filtering
70•harperlee•4d ago•4 comments

Trillions spent and big software projects are still failing

https://spectrum.ieee.org/it-management-software-failures
139•pseudolus•7h ago•124 comments

It is ok to say "CSS variables" instead of "custom properties"

https://blog.kizu.dev/css-variables/
55•eustoria•2h ago•39 comments

Show HN: Secure private diffchecker with merge support

https://diffchecker.dev
9•subhash_k•45m ago•6 comments

Human brains are preconfigured with instructions for understanding the world

https://news.ucsc.edu/2025/11/sharf-preconfigured-brain/
362•XzetaU8•13h ago•247 comments

Jakarta is now the biggest city in the world

https://www.axios.com/2025/11/24/jakarta-tokyo-worlds-biggest-city-population
81•skx001•13h ago•27 comments

Making Crash Bandicoot (2011)

https://all-things-andy-gavin.com/video-games/making-crash/
158•davikr•7h ago•16 comments

LPLB: An early research stage MoE load balancer based on linear programming

https://github.com/deepseek-ai/LPLB
16•simonpure•6d ago•0 comments

Ozempic does not slow Alzheimer's, study finds

https://www.semafor.com/article/11/25/2025/ozempic-does-not-slow-alzheimers-study-finds
68•danso•3h ago•46 comments

Most Stable Raspberry Pi? Better NTP with Thermal Management

https://austinsnerdythings.com/2025/11/24/worlds-most-stable-raspberry-pi-81-better-ntp-with-ther...
254•todsacerdoti•13h ago•79 comments

IQ differences of identical twins reared apart are influenced by education

https://www.sciencedirect.com/science/article/pii/S0001691825003853
16•wjb3•22m ago•8 comments

Unpowered SSDs slowly lose data

https://www.xda-developers.com/your-unpowered-ssd-is-slowly-losing-your-data/
673•amichail•1d ago•277 comments

Roblox is a problem but it's a symptom of something worse

https://www.platformer.news/roblox-ceo-interview-backlash-analysis/
122•FiddlerClamp•3h ago•174 comments

Inflatable Space Stations

https://worksinprogress.co/issue/inflatable-space-stations/
18•bensouthwood•4d ago•6 comments

Broccoli Man, Remastered

https://mbleigh.dev/posts/broccoli-man-remastered/
125•mbleigh•6d ago•69 comments

Orion 1.0

https://blog.kagi.com/orion
217•STRiDEX•3h ago•122 comments

PRC elites voice AI-skepticism

https://jamestown.org/prc-elites-voice-ai-skepticism/
92•JumpCrisscross•23h ago•25 comments

US banks scramble to assess data theft after hackers breach financial tech firm

https://techcrunch.com/2025/11/24/us-banks-scramble-to-assess-data-theft-after-hackers-breach-fin...
38•indigodaddy•2h ago•0 comments

Nearby peer discovery without GPS using environmental fingerprints

https://www.svendewaerhert.com/blog/nearby-peer-discovery/
55•waerhert•4d ago•17 comments

Claude Advanced Tool Use

https://www.anthropic.com/engineering/advanced-tool-use
617•lebovic•1d ago•247 comments

Brain has five 'eras' with adult mode not starting until early 30s

https://www.theguardian.com/science/2025/nov/25/brain-human-cognitive-development-life-stages-cam...
218•hackernj•6h ago•195 comments

APT Rust requirement raises questions

https://lwn.net/SubscriberLink/1046841/5bbf1fc049a18947/
217•todsacerdoti•5h ago•387 comments

Using an Array of Needles to Create Solid Knitted Shapes

https://dl.acm.org/doi/10.1145/3746059.3747759
72•PaulHoule•3d ago•20 comments

Hollywood's vision of ancient Rome is all wrong, according to Mary Beard

https://www.openculture.com/2025/11/why-your-vision-of-ancient-rome-is-all-wrong-according-to-his...
59•bookofjoe•6d ago•57 comments