Hacker News new | past | comments | ask | show | jobs | submit login
macOS Free and Open-Source Security Tools by Objective-See (objective-see.org)
402 points by behnamoh on Sept 26, 2022 | hide | past | favorite | 48 comments



These tools send telemetry to Sentry (or at least the ones I evaluated did, which I believe were Lulu and TaskExplorer).

Personally I'm not comfortable with this due to the level of access they require and thought it worth to mention.


Lulu setup for Sentry seems to be for crash dumps being sent:

https://github.com/objective-see/LuLu/blob/dfe48c2a0f7a86361...

It does seem like a good idea that this could be disabled but now it's hard coded.

Issue on the topic: https://github.com/objective-see/LuLu/issues/488


This particular line in the developer's response[1] leads me to believe they don't - or at least don't properly - validate third party dependency changes, which is even more worrying if I'm honest:

  > I will look into why Sentry.io is being contacted so often! This is strange as unless they changed their SDK/framework, is only supposed to happen on a crash report.
1: https://github.com/objective-see/LuLu/issues/488#issuecommen...


That's a poor characterisation, I'm not convinced any developer can say with 100% certainty they've never missed anything in the release notes of their dependencies or had any unexpected behaviour in changes that made it to production. Mistakes happen.


The nature and capabilities of these products makes me significantly less forgiving towards these sorts of mistakes, however.

Just to be clear I have no ill will towards this developer or their products.


I agree with sibling: I think you're giving it an uncharitable interpretation. There is an open issue and they are investigating to improve the situation.


You’d think, if they wanted telemetry so badly they are prepared to deal with the fallout, they’d be on top of the information collected by that telemetry.


> crash dumps being sent

Crash dumps can contain highly sensitive info. I have blocked Crashlytics[0] and Sentry with my pi-hole so nothing gets sent. I'm starting to believe these tools are abused and not even used to improve services. They're basically a MITM so people can violate other's privacy.

[0] https://firebase.google.com/products/crashlytics


Can the same be done with hosts file (block crashlytics/sentry?) for those of us unable to use pi-holes at work?

edit: it looks like it is working, but unknown if there are any negative side-effects at this point.

    └─[$] <> ping sentry.io
    PING sentry.io (0.0.0.0): 56 data bytes
    ping: sendto: Socket is not connected
    ping: sendto: Socket is not connected
    Request timeout for icmp_seq 0
    ^C
    --- sentry.io ping statistics ---
    2 packets transmitted, 0 packets received, 100.0% packet loss


You can block the root, but usually the domain contains random (unique) strings as a subdomain, so you need to wildcard block it, which pi-hole supports.


Thanks. Bummer, I figured they probably didn't use the root domain but some amalgamation of letters/number/sub-domains. I have a pi-hole at home which has been working great for 3+ years now, but I mostly run objective-see stuff on my work machine.


You can block Lulu with Lulu. I think by now we should all have learnt to block internet access for every app, and only enable it if it really needs it - that's exactly what an application firewall like Lulu is for.


Potentially relevant that the Sentry client is configured in debug mode there. Not sure what the actual impact of that is but maybe it causes additional network traffic to Sentry (outside of crash reports)?


you can actually use lulu itself to block sentry for all of their tools (this is what i do). i found that the only other connections they'll make is for updates--both software and rulesets (if you have that configured).

lulu is not nearly as polished as hands off! (may it rest in peace) or little snitch, but it gets the job done. i also have some rules in pf (via murus gui) to block things like google and facebook on a system-wide level.


It's troublesome that an app designed for blocking telemetry (among other things) is sending telemetry without consent. :(

Software that exfiltrates your usage, crashes, or other data from your own machine without advance, opt-in, informed consent is unethical and disrespectful.


I sent a message to slack the other day recommending Shortcat (https://shortcat.app/). I edited the message to mention that I use LuLu and it's worth noting that Shortcat is closed source and does send info out to sentry.

Now I have to edit that same message and mention that LuLu also phones home to sentry. Can't blame people for wanting stack traces but wow it's a tricky subject in terms of privacy.


Thanks for the heads-up.

I've been using Shortcat for years, and I thought it was long since abandoned. I was even more surprised to see a new version was released last month. I seriously stopped checking on this project years ago, and just kept a backup .dmg.

I am sad to read about sending out info to Sentry, but I guess that is something I am going to have to think about some more.


I wish LuLu had some kind of logging though, and asynchronous rule adding (i.e. no need to take decision when connection happens and block by default, and then review this list when I have time), otherwise it at some point you just start accepting connections for all processes without thinking, just to not be distracted too often.


Perhaps a good start would be to habitually deny connections instead of allowing them. Otherwise the need for Lulu is rather unclear.


I’ve been using KnockKnock, RansomWhere?, BlockBlock, and OverSight for many years. They are great pieces of software that are simple to use and do what’s advertised. The biggest issue I’ve had is memory consumption issues with OverSight.


Some of the best documentation I've ever seen. Click on the "Learn More" for any of the apps (I suggest ransomwhere). Other software developers should take note that this is how it should be done.


> Objective-See

Heh, I C what you did there.


Seems you can C #.


How is LuLu compared to littlesnitch? And is it even useful anymore with macOS being increasingly locked down (see https://news.ycombinator.com/item?id=24838816).


> How is LuLu compared to littlesnitch?

Last i tried LuLu it worked similar to Littlesnitch, though nowhere near as "polished". The basic functionality is more or less the same though, with Littlesnitch offering to automatically unblock known "trusted" services like Apples own services.

> And is it even useful anymore with macOS being increasingly locked down

It's an application firewall, so even if MacOS is locked down, any app can still roam freely (within it's jail). Suppose some app has access to your contacts, that means it can still upload every contact to a server, and an application firewall can help you detect/block that.


Blocking the Apple services that every mac incessantly phones home to (with unique identifiers tied to hardware) is the main use case of Little Snitch for me personally.

Even if you don't use iCloud, the App Store, iMessage, FaceTime, or any of it - macs still send tons of realtime usage data to Apple even if you don't want them to.


This was fixed almost immediately after this came about.


LuLu you can build yourself. Both protect you from 3rd party apps, not Apple, as Apple can do whathever the ** it wants to your computer through updates.


False. Little Snitch can block connections from all processes in the OS, including the Apple factory ones.


You can't disable automatic updates?


You can.


KnockKnock is pretty useful. I run it once a month to make sure nothing out of ordinary is installed on my Mac.

I once found out that a VPN that I have uninstalled long time ago still has shady entry in the login items thanks to KnockKnock.


Hate to say it, but I used to use their tools, but since moving to little snitch (and micro snitch), I don't really have a need for them.

Some of the tools are still useful to have installed though.


Little snitch is top notch software. Can’t recommend enough.


I'm using LuLu since few months and find it absolutely helpful.

When I first installed it, it required quite some efforts to consciously filter/allow traffic from/to for the apps. By the time, all regular apps were properly configured for the rules and now I see notifications for block/allow only when there is uncertain traffic going out.

Definitely recommend to give it a try.


Very cool! Does anyone know of such OSS tools collection for Windows? Desperately looking for some good solution for a server I run for a small office.


What you can do is run Windows through Proxmox.

On Linux, I use OpenSnitch. The name's based on LittleSnitch. Its a layer 7 firewall. Quite polished these days.



It's not Open Source, just freeware. I would not trust anything not Open Source to filter my connections.


Somehow the code is not linked in the blog posts, but it seems to be there: https://github.com/objective-see?tab=repositories


Not open source comment was for Nirsoft, not about Objective-See (and I just checked, source code link is present on home page as well as individual articles)


I am working on this, https://www.appswithcode.org/?q=windows , It is far from ready, but it is meant to solve such problems.


Ohhh i've never heard of whatsyoursign - awesome list. cheers


When will websites stop asking me to sign up for a newsletter before I've even seen the content? I closed immediately. Just honest feedback


I know it's a rhetorical question. An answer could possibly be:

1. when our "browsing, and information self-exposure" tools are better (automatic note taking parrot robot that sits on your shoulder and remembers everything you've seen so you don't have to) and

2. when our Internet's base concepts are more equitable to content creators/intellectual property owners.

More:

For number one, obviously it's handy if you're interested in a website to be reminded of that website latest and greatest successes.

Number two, with the Lamina1 news recently it's got me thinking again about the inequitable economy of providing useful advances and information for free, or in this case tools, and then not being respected by the world in a way which the pressures of reality direct you to collapse or shut down your fantastic enterprise .. again in this case of creating macos anti-malware tools.

(Social comment: I see identifying a UX problem is one step in responding to someone's work, and the ramification of talking about your frustration is another. There's at least one more you can do, call to action: how would you, Message Poster with the beef against that UX, have offered to solve, or make better, these problems responsibly if you were the owner of the website?)


Not OP, but have similar behavior. Maybe I'm overthinking things, but I think you're blowing the gripe out of proportion.

Showing me a newsletter signup before I've read the content implies that I'm interested in getting more of what I expect the content will be about, not what it actually is. Asking me to sign up for more interesting sounding titles before I've even had a chance to decide if I enjoy the content within implies that you, the content author, don't actually care if I enjoyed the content. What you're most interested in is pushing more clickbait titles in my face.

Put a newsletter sign up button near the end of the content, or in a side bar next to the content- anywhere that makes it seem like I'll get more of the content I am enjoying.


When it stops being a lucrative pattern for harvesting emails. Unfortunately.


When the "engagement team" gets penalized for closed browser windows.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: