Hacker News Reader

Comment by supernova87a | Hacker News Reader

Why are people out to crucify Apple for a story that's still being resolved? The article clearly says:

"...Due to a processing issue, your credit will be included on the security advisories in an upcoming update. We apologize for the inconvenience," Apple told him when asked why the list of fixed iOS security bugs didn't include his zero-day..."

"...We saw your blog post regarding this issue and your other reports. We apologize for the delay in responding to you," Apple told Tokarev 24 hours after publishing the zero-days and the exploit code on his blog...

"...We want to let you know that we are still investigating these issues and how we can address them to protect customers. Thank you again for taking the time to report these issues to us, we appreciate your assistance..."

The company hasn't denied the bounty, they're just incompetent / slow on this process.

Feels like everyone is out to paint <x> company with just confirmatory bias using whatever half-baked story is available. Even I feel for company leaders in this kind of shitty journalism environment. And the rest of the comments here are just autopilot piling on the echo fest.

View on HN

You're framing this as if it's all about the bounty and Apple just hasn't gotten around to it yet. That's only a small fraction of the story and could easily be forgiven. If I wanted to make Apple look good, I'd focus on that part, but that would be rather biased to ignore the whole picture...

Tokarev discovered 4 iOS 0-days, then reported them all to Apple back in May. After months of Apple's continued refusal to fix or even publicly acknowledge all four of the issues, Tokarev made all of them public on GitHub.

Weeks passed, and now it's today. Apple has yet to fix or publicly acknowledge two of the four security vulnerabilities. That should make Apple look bad because it's some fundamentally irresponsible security practices.

Yes, I'm biased. I'm human, not a computer, and it's stuff like this that makes me biased towards Apple. They should receive negative publicity for this, then they should change how they do things. At the very least, app developers and users should be warned about the two issues that have yet to be fixed.

I don't know anything about Apple's bug bounty program except that there's a prevailing attitude that it is not the well-oiled machine that Google's bug bounty program is perceived to be, and I'm not super interested in making a case for Apple here. But because this is a recurring theme in every discussion about every bug bounty run by anyone:

* There are valid reasons that bugs can take longer to fix than you'd expect; the most notable of them is when the bug you found is actually systemic, or has a deep root cause, and the real fix for the vulnerability is more complicated than the surface bug. Without a hard timeline, some shops will work to get the root cause fixed on some bugs even at the cost of an increased timeline, because the patch for the surface bug reveals the pattern and amplifies risk to customers.

* As a reporter, you can take some measure of control over the process back by providing a fixed timeline (like the P0 90 days). There's no negotiation needed; you give the vendor time to fix and they either do or don't, but either way you're going public. That is a valid way to go about things, but may cost you the bounty.

* These things are bug-dependent, and the process that runs for a zero-interaction RCE won't be the same as the process that runs for a bug that requires a malicious app store app and only gives access to the contact database.

* Message boards tend to expect that big vendors can just shell out for the bounty as a show of good faith. It's easy to see why they believe that. It makes sense. But it also creates broken incentives. The limiting reagent on bugs isn't bounty dollars (these are indeed barely even rounding errors to major vendors), but rather programmer time. If you pay out for weak, stuck-in-process bugs, you create incentives that redirect programmer time to those weak bugs and away from more significant bugs; as angry as you can reasonably be about a malicious app being able to snarf your contacts, if you're rational, you're a lot more concerned about memory corruption flaws, which is what you really want people spending their time on.

> If you pay out for weak, stuck-in-process bugs, you create incentives that redirect programmer time to those weak bugs and away from more significant bugs; as angry as you can reasonably be about a malicious app being able to snarf your contacts, if you're rational, you're a lot more concerned about memory corruption flaws, which is what you really want people spending their time on.

I don't understand how this isn't already reflected in bug bounty pricing tiers? Like, if I access your contacts, I get $x, but if I can access all your photos, I get $xx. As a user, I couldn't care less for how my data got accessed, whether it's a logic bug or memory corruption…

The point is that memory corruption vulnerabilities are complete device takeovers, not that they have extra aesthetic value.

Right, and responder is saying in a competition for resources, pay less for non-mem corruption bugs, which should mean they aren't as hard to verify.

Is your point that all bugs soak up time to fix? If so pay the bounty and add it to the backlog. Or is it too much time to verify? That seems to be something you can kick back to the reporter.

Apologetics for a bug bounty program dragging their feet on a payout because "it's not a mem corruption bug" alone is unconvincing.

Well, that circles back around to my point: if Apple wants people to submit more complete device takeovers, they should pay more for them (and they do, or at least they claim they do).

This bug was not a complete takeover, or anything even close to one.

Well, yes, that's why Apple won't pay a million dollars for it. But there are tiers lower than that ("Unauthorized access to sensitive data from a user-installed app") which do cover the kind of bug described here.

You've lost me. What does this have to do with the point I made about which bugs we want security developers at Apple spending their time on? You seem to be making my point for me.

The whole point of the bug bounty program is that anything that qualifies for it is something that Apple considers to be important enough to pay out money for, because they (ostensibly) want people to report these to them. I understand if they are overloaded with "this function leaks the country of the device" bug reports but the ones that they actually include in the bug bounty program are those that have the potential to significantly damage their stated goals around security, so it's kind of the point that they review all of them.

If Apple wants to nudge people towards submitting more serious bugs, they can pay more so people are incentivized to work towards those rather than mucking around in gamed. But, they don't really get to say "we didn't really have time to get around to this Contacts bug, sorry": people expect them to have it fixed. That's the whole reason Apple offers a bounty at all: so that researchers tell them about it early so they can fix it.

I think there might be someone else on this thread you mean to direct this to, because it's not responsive to anything I'm saying.

> * These things are bug-dependent, and the process that runs for a zero-interaction RCE won't be the same as the process that runs for a bug that requires a malicious app store app and only gives access to the contact database.

It would be interesting to understand at what point this becomes a GDPR issue, and if the GDPR legislation can be used to pressure companies in expediting this process.

There's a "be careful what you wish for" argument here, because my understanding is that the FAANG vendors are snapping up security people just as fast as they possibly can, and, again, ceteris paribus you'd rather have those people working on the actual most serious vulnerabilities rather than the ones causing the noisiest bounty drama. But you could reasonably go either way on this I guess.

The degree of damage done by particular vulnerabilities is different from person to person. For one individual losing their contact database to some un-identified third party might be a 'meh' event, for another it could be a disaster.

Devil's advocate here: I've worked the other side of managing bug bounties.

It is entirely possible the researcher found something but didn't realize how deep the problem went. Apple may have released an incremental patch and is working on fixing a larger issue they found when digging into it.

When this has happened in the past, from the researchers perspective things seem quiet/delayed because we obviously can't share details of a larger vulnerability with them. All we can really do is ask for more time. In the end it all works out and they get paid out/credited for the original+follow on bug.

Why wouldn't the company communicate to the researcher "we found a larger issue related to this. your bounty will be upgraded to X. Please restart the clock for public disclosure" or something along those lines. Seems like better communication would create a win-win situation.

My first thought would be that the team within Apple may worry the researcher may resell the vulnerability to an exploits site.

Not part of the security industry so not sure what is common or not, but I would understand Apple being worried about sharing too much with a researcher they may not be familiar with.

I would also understand the researcher's point of view that this fell through the cracks or Apple is not willing to fix; and is likely what happened.

Apple should be paying enough money that that issue is not a consideration. If I’m Apple (or anyone else for that matter) I’m paying absolute top dollar times two to resolve these issues. And I’m not even thinking twice about it.

that is, unfortunately, not at all how the bug-bounty market works. Apple (or any other tech company) can't outbid three-letter-agencies, certainly not on a regular basis. Open market value is at least 10x higher than companies will pay directly.

Apple will pay a million bucks? Fine, NSA TAO will pay $10m. Apple can't pay $10m or $100m a bug on a regular basis, for the customers whom this matters the check is basically blank, as much as it takes.

How does one contact the NSA TAO and offer to sell a zero-day?

> We pay big bounties

https://zerodium.com/

>One person who will share those sales numbers is a South African hacker who goes by the name “the Grugq” and lives in Bangkok. For just over a year the Grugq has been supplementing his salary as a security researcher by acting as a broker for high-end exploits, connecting his hacker friends with buyers among his government contacts. He says he takes a 15% commission on sales and is on track to earn more than $1 million from the deals this year. “I refuse to deal with anything below mid-five-figures these days,” he says. In December of last year alone he earned $250,000 from his government buyers. “The end-of-year budget burnout was awesome.”

https://www.forbes.com/sites/andygreenberg/2012/03/21/meet-t...

For those who figure this is a great way to monetize their security skills and actually have the chops to do it:

It should probably be pointed out that once you do this, you’re in the weapons industry. Your work will likely be used, directly or indirectly, to put a bomb through someone’s roof or put them in prison for a very long time. Make sure you’re okay with the ethics of it.

Getting rid of people who want democracy

> It should probably be pointed out that once you do this, you’re in the weapons industry. Your work will likely be used, directly or indirectly, to put a bomb through someone’s roof or put them in prison for a very long time. Make sure you’re okay with the ethics of it.

By this logic, americans should stop using cars at all, cause all that oil is coming from middle east, saudi arabia.

What? This does not follow at all. You’re implying that any degree of involvement in activities that have negative consequences is equivalent. That’s incorrect!

American cars mostly run on gasoline from domestic and Canadian sources. Very little comes from Saudi Arabia.

One doesn't, I don't think; I think one sells to one of several grey-market brokers who in turn sell to DOD. But I think it's more productive to substitute "the IC" for "NSA TAO", because there are several countries (on the "sort of legitimate" side of this market) buying. All of them can pull any plausible amount of cash for a vulnerability out of their couch cushions (then again, so can small countries).

They can’t be that worried - the bug that allows you to input behind the lockscreen on OS X has now been a thing for… six years, over three or four versions of macOS? I’ve reported it, but they each time said it was a feature, not a bug.

Damn strange feature that allows me to compromise any screen-locked mac.

How do you do that?

Surprisingly straightforwardly - when you open the lid, focus is initially still on the desktop, not on the Lock Screen. You may have noticed if you’re too quick with your password the first few characters don’t appear. They don’t go nowhere - they go to the desktop.

If you get a certain key combination in before focus switches, it stays on the desktop, and you can continue to input - fire up a terminal, do whatever you fancy. It’s all blind, but still perfectly dangerous.

Certain full screen apps, if they have focus when you lock, retain focus indefinitely. Paradox interactive games, for instance - stellaris exhibits the behaviour nicely. This even includes mouse focus, and if you Apple-tab, then focus goes to whatever you Apple-tab to. You can only restore focus to the Lock Screen by clicking in the password field.

Both times I reported this I got a pedantic “locking the screen does not terminate applications, which may continue to run in the background” response.

Open-terminal shortcut, then "killall xlock" or whatever it's called in macOS?

I would assume that trying to get a bounty means accepting that you're not selling the vulnerability elsewhere. No maybe the shadiness inherent in the field means no one trusts that agreement anyway, but I suspect there's generally some amount of trust for researchers submitting vulnerabilities, particularly if they have some kind of history of good faith.

Also, I'm not sure that saying "we have discovered a deeper problem that here beyond what you reported" really delivers much information beyond perhaps telling the researcher to keep investigating (although if they're already getting the bounty, the additional investigation wouldn't really be useful).

> would assume that trying to get a bounty means accepting that you're not selling the vulnerability elsewhere

Having an e-mail from the company confirming the bug is serious and systemic massively raises its market value. Security is necessarily trust less. These game dynamics are unavoidable.

Having an email from the company confirming that they know about the bug probably erases its market value.

Unless it's harder to fix, and easy to short term exploit?

None of us are thinking of any of this before the brokers already have; this work has been going on for many, many years. You can resell vulnerabilities, but the contract terms you get from brokers are tranched, and they stop paying when the vulnerability is burned. Your incentive not to spread your bug around is that you you're cutting your nose off to spite your face, not to mention that you're probably making the terms you'll get for future bugs worse.

> I would assume that trying to get a bounty means accepting that you're not selling the vulnerability elsewhere.

That is a very mistaken assumption. Even NDAs backed by threats from nation state intelligence agencies aren’t sufficient to keep exploits from being resold multiple times.

Also, the ones who sell two or more times, can just claim that the other buyer actually discovered the exploit themselves?

Or maybe that someone in the first buyer's organization resold it?

> In the end it all works out and they get paid out/credited for the original+follow on bug.

I've worked on the company end of bug bounties too, and it does happen that a report just falls through the cracks. Seemingly-inactive reports do need a certain amount of maintenance; you don't want to just trust that everything will work out in the end. (That said, as long as you get responses when you ping the company, things are working in the background.)

(edit to followup: in about 18 months of this, I encountered one report that had fallen through the cracks. Obviously, there might have been others that never came to my attention at all, but the companies are tracking things much more carefully than researchers often assume.)

I have also worked on managing bug bounties and that is why you keep lines of communication open with the researchers. Not to throw stones in glass houses, but there are a number of ways Apple could improve on their approach to how they do their bug bounty program.

I have heard of many researchers having extremely long delays, poor communication and simple things like not acknowledging the bug submissions.

Aftering filing multiple bug reports with Apple now against WkWebView, I can say that extremely long delays and poor communication is not isolated to the bug bounty team :)

Security bugs are much more important than other bugs. A security bug that's affecting no one currently could turn into a disaster tomorrow. A functionality bug that's affecting few people today will almost certainly affect few people tomorrow.

Why withhold payment though?

"Since then, Apple published multiple security advisories (iOS 14.7.1, iOS 14.8, iOS 15.0, and iOS 15.0.1) addressing iOS vulnerabilities but, each time, they failed to credit his analyticsd bug report."

"Two days ago, after iOS 15.0.2 was released, Tokarev emailed again about the lack of credit for the gamed and analyticsd flaws in the security advisories."

They didn't give him credit in the last 5 advisories. Really no excuse for that imho. If Apple keeps this up then why would anyone report bugs to them when you can just post it online and get credit for it right away? Or sell it on some 0-day site.

If credit is what you care about, it's straightforward to ensure you get credit without working with Apple's bounty program. You can do what P0 does and provide a fixed timeline after which you're publishing, and nobody credible is going to hold that against you (in part because P0 has established this norm).

Obviously credit is important for them as a proof of competence. If the company does not give them credits how can they build their business, portfolio. You can jus say I was the one who discovered this.

Every field works in a certain way and when it comes to bounty you want to make a name for yourself. You can't just pull up and say you are the one

But yeah may be they should just sell it to third-parties

Exactly, that's the issue. This is a RT*A scenario.

> they're just incompetent / slow on this process

> I feel for company leaders in this kind of shitty journalism

You're not making sense. Plus Apple has a history of being incompetent and slow on this.

>Plus Apple has a history of being incompetent and slow on this

Have they, really? Just because you find this instance here and there of such a story where they were, doesn't mean they have a history of being incompetent and slow on this (the same way someone who hit 99% of their three-pointers doesn't have a history of being an awful shooter).

That's how they fare long term:

https://www.pandasecurity.com/en/mediacenter/mobile-security...

Apple's bug bounty is notoriously slow to respond, to the point that it has posed legitimate security concerns in the past: particularly their rhetoric around Thunderspy amused me.

https://habr.com/en/post/579714/

https://thunderspy.io

Is this not a potential Sharpshooter fallacy? Are you sure that we aren't mainly hearing about bounties when Apple doesn't handle them smoothly and not so much when they are handled smoothly?

Or to put it another way, when you poll people who complain Apple is notoriously slow to respond, don't be surprised if your conclusion is that Apple is notoriously slow to respond.

How can we reasonably say Apple is moving slowly to fix bugs when we don’t know how much work is going on behind the scenes? A slow response can just be a slow response.

I can reasonably say it because all of their competitors have less cash, yet many of them respond faster than Apple does.

Probably because the 0days have been fixed and the public have received the benefit of the security researcher's report to Apple, but he hasn't received any recognition or compensation through the bounty program via which he reported the vulnerabilities?

>Why are people out to crucify Apple for a story that's still being resolved?

>The company hasn't denied the bounty, they're just incompetent / slow on this process.

People probably expect more from... checks notes The world's most valuable and successful modern corporation.

And iOS users should be grateful that they report those bugs to Apple to get paid. They could also sell it to some spying companies and may be those pay well and very fast

> and may be those pay well and very fast

Pay very well? Often, assuming they actually pay, sometimes you can get stiffed there too. Very fast? Nope. Easy to work with? Nope. Communicate with you any better than Apple through the process? Not usually.

Have you dealt with terrorist orgs in the past, or is this all conjecture?

> They could also sell it to some spying companies

The above type of organization is what I was referring to. So if you consider the grey market buyers[⋁] of exploits “terrorist orgs”, then yes. If you use the normal definition of “terrorist orgs”, then hell no.

[⋀] three letter government agencies, defense contractors, and those who sell to them

Or maybe they don't pay for these bugs at all.

Ya, if I interpreted this right, also really convenient that they seem to be dragging their feet on a $100,000 bounty.

Nobody's going to pay $100,000 for a bug that lets you download someone's contact list.

How much would a bug like that be worth? I can imagine getting anyone's contact list being valuable.

Why? What would be valuable about that? Specifically, what would be valuable about getting someone's GameCenter contacts? These aren't business or family contacts. These are people the person games with. It's a privacy violating bug but not a show-stopping bug. Important but not valuable.

You're misunderstanding the vulnerability. The bug is in gamed, the Game Center daemon, but it allows access to the entire CoreDuet database, which does on-device intelligence stuff. Duet essentially logs everything you do on your phone, which means that if you look at the database it'll contain logs for all your interactions, not just those with Game Center contacts.

You are correct. I did misunderstand that portion of it but that doesn't change the fact that Apple is willing to pay $100k for a bounty like that but no one else does and Apple's actions don't suggest at all that they won't pay out the bounty. If you've ever been part of a program like this, on the developer side, it's possible that they discovered a larger bug that this was a part of or that someone else had already reported this bug. Tracking down the origin of this and the tickets involved takes time and fixing the bug is always the priority. They haven't acknowledged that he's the originator of the report so his insistence that he be credited immediately is a big immature and premature.

There's nothing here that suggests Zerodium, or someone similar, would pay the same amount Apple is offering and there's nothing that suggests that Apple doesn't intend to pay this or credit him. That's all completely conjecture.

I have no idea if Zerodium would pay out that much for this bug (probably not) or anyone else (maybe?) but Apple in general has a poor track record of paying out bug bounties. They do sometimes, but it seems like it is far rarer than their website says they should.

>Apple in general has a poor track record of paying out bug bounties

What is this based on? My understanding is that Apple pays out 99% of the reported bug bounties and that's only because they include multiple submissions in the totals but not in the payouts (they only payout the first discovery or root discovery).

https://habr.com/en/post/579714/

thunderspy.io/

Those are my favorite recent examples, but specifically Apple has huge issues with turnaround time. They also don't communicate with or assist the researchers who found these exploits either, which makes things particularly frustrating for people who ultimately both want to secure Apple's systems. Their overt hostility, history of poor communication, and frankly pathetic bug bounties are all contributors to how people perceive Apple's relationship with security experts.

You just posted the article that this whole thread is based on. You just posted the author's post instead of the blog's rewrite of it.

This makes it seem like this is a recurring problem yet there are only a handful of complaints.

> The company hasn't denied the bounty, they're just incompetent / slow on this process.

With 0-day security vulnerabilities, slowness equals incompetence. Companies with unbounded resources like Apple have absolutely no excuse for not being able to move as quickly as a small startup on issues like this, unless their message to shareholders is "yes invest in us so you can see our performance literally decrease with every dollar invested".

Let me selectively quote you:

> Why are people out to crucify Apple (...), they're just incompetent / slow on this process.

That's exactly the problem when they're endangering the security of approximately one billion users.

Some security people love to hand-wave and issue prophecies of doom for attribution and attention. It’s great chum for writers — easier to run with some guys grievance than research a more substantive story.

There's a spectrum of quality to these stories, and one sign that you're tending towards an end of that spectrum is the use of the term "zero-day" without qualification. These are bug bounties; all of these bugs are zero days, no matter how severe (or not) they are. It's literally the least important detail in the story about how a bounty is being handled.

Are there lots of articles that describe zero-days but don't include the actual term "zero-day"? I may have been underestimating the state of journalism...

The “Apple Hater” market is as big as the “Apple fanboy” market, maybe even bigger!

Because those emails are probably lies, they're just delaying and delaying and hoping it will just go away, and writing fake "We're sorry"'s when they're forced to.

Reply on news.ycombinator.com