side note but what is the consensus on figuring out whether small brew packages like this one are malicious?
I'm sure this one probably isn't, but I have no clue how I would know whether it is without reviewing the source code myself (or pointing an agent at it etc)
Even if they're not malicious, there's no guarantee that they won't become malicious in the future. Prefer living off the land, and install only the very essential extras.
Good question -- I have been pointing claude at repos for security reviews before I install stuff now for exactly this reason.
It's so funny to read these, as if countless examples didn't prove that llms don't actually distinguish between your instructions and the data they read at your request.
Do you think a malicious repo author won't think of hiding a suitable prompt within?
> countless examples
Show me just one where a frontier model got prompt injected when checking if code if malicious.
FWIW I can do this, but I bet you can't.
Here's prompt injection via a harness for example: https://www.promptarmor.com/resources/google-antigravity-exf...
took me about 30 seconds to find. there's tons of them. of course it's an arms race, the models get better the harnesses get better the attackers get better but again: for the model itself, your prompt and third party data it reads are the same thing (input). as long as it is true, the attack will always be possible.
Basically the question for any executable code (or even just any data) that you let on your machine and which did not go through independent vetting processes. The question you need to ask is if the benefit (getting rid of an annoying popup) is worth the risk of a distribution method that is trivial to compromise.
so we are back at choosing where we draw the line and you picked this one to get us going
That's why I like small, open source projects like this one. I don't have to figure out if it's malicious. I just fork it, then ask claude to review the code, then run my own fork.