Hacker News new | past | comments | ask | show | jobs | submit login
Dangerous bug in Chrome’s ‘New Tab’ page bypassed security features (portswigger.net)
136 points by PaulHoule on Nov 18, 2021 | hide | past | favorite | 91 comments



I dunno what audience this article is aimed at, but it could do with trying to use less abbreviations - I've never seen the new tab page be abbreviated to NTP (that's the time server thing after all).


Big companies tend to do this sort of thing, they are large enough to ignore convention 'outside' and it tends to give the insiders the feeling that they are special, it's another form of gatekeeping. You see the same in the military with endless acronyms.

On a smaller scale, tech people do the same thing by using more complex terms for simple things to appear to have some kind of special knowledge. It's all about who is on the inside and who is on the outside. Highly annoying.

Such DSLs can serve to increase the speed of communication but more often than not they are simply used for obfuscation purposes.


> Such domain specific languages can serve to increase the speed of communication but more often than not they are simply used for obfuscation purposes.

(I think I missed the self referencing joke).


Hah, good one, I just commented in a thread about Forth and still had that acronym on the brain, but you are totally right, that's a perfect example.


> You see the same in the military with endless acronyms.

The military take the abbr.hl. to the next level. But atleast the abbreviations are properly documented there. I guess the root is keeping telegraphy short?

On my last job it was so bad that it took like a year before you could follow conversations properly. Also old deprecated abbreviations were used for extra flavor. E.g. calling projects or departments by their former former name.


My current employer is similarly terrible in this area. It took a year before I felt like I fully understood people’s day-to-day conversations.

On top of the usual acronym madness, nearly everything is always referenced by code name, versioned according to arcane and strange conventions, and the mapping to released product names and real version numbers is not always documented or obvious.


From the bug report it seems the chromium team uses the NTP term: https://bugs.chromium.org/p/chromium/issues/detail?id=125154...


To be fair to the article, I could only find NTP in that use as part of a quote. But I agree – it’s an overloaded initialism.


When will it be possible to have a New Tab page which doesn't display anything?

Just kidding, trick question, the answer is NEVER.

(No, an extension which overrides all the bloated pile of crap after it's already been processed and rendered does not count.)


> Just kidding, trick question, the answer is NEVER.

What? The answer is ALWAYS. Set your start page to be `about:blank` and you see a blank page. I've had this as my starting page in every single browser since the 90s.


Sorry, but this option is not available at least as far back as Chromium 65.x


You can have about:blank be the on start up page and the home button, but not for new tabs. You used to be able to do that. Won't be surprised when start page will need to be an https link, you know for security reasons.


As a side note, only yesterday (on a Windows 7, so not an issue related to the latest Windows) I couldn't initially connect with Chrome to a (of course local) oldish router (actually an access point) to change a setting (Wi-Fi channel) because it "talked http" while Chrome wanted a "https" (for security reasons).


That's "start page", not "new tab page"


I am not able to in Chrome Stable without an extension. Searched chrome://settings for "tab" and "page".


I have that since ages in Firefox. A simple setting.


Chrome has that setting too, but it's not that simple. It's only available as an enterprise policy. If anyone doesn't know about this, Chrome has tons of hidden settings configurable through Group Policy on Windows and through /etc/opt/chrome/policies/managed/policies.json on Linux.

https://chromeenterprise.google/policies/#NewTabPageLocation


For those not in the know: it's a setting the GUI settings (so not just in about:config somewhere). So it's not that hard to find.

Preferences » Home » New Tab page [Firefox startpage v]

or, about:preferences#home


It's available in Brave too


> No, an extension which overrides all the bloated pile of crap after it's already been processed and rendered does not count

afaik, extensions that set chrome_url_overrides.newtab in their manifests file prevent the native NTP from loading at all.


I have a personal new tab chrome / firefox extension that does exactly this. A black screen, a button, that's it. Don't install someone else's extension -- make your own off a minimal example on github. It's... well it's about the simplest bit of code I've written that I rely on daily.


Use a different browser? It's one of two built-in options in Firefox.


I generally do, but sometimes I use Chromium, if only for testing my own sites.


Can't you just set the new tab page to be "about:blank"?


Not in chrome. You can only set the startup page to be about:blank. New tabs do not follow that startup page setting.


Firefox can do that, for Safari you can also disable all start widgets. IIRC Brave can do it aswell.


Firefox allows it through settings. Haven't been able to do so on Chrome, other than disabling "NTP Modules" via chrome://flags/.


u serious?

i would never use a browser w/o a blank new tab page; it's universally supported isn't it?


The first thing that I do is turn off these fancy new tab pages. Very often, there is no option to have a blank page instead and less and less people know about pages like about:blank


Firefox is making it real easy to have new windows and tabs open about:blank.

For all the hate they get, FF is easily the best and most respectful browser around


The dislike (not hate, at least in my case) is not against Firefox. I like Firefox a lot.

What I don't like is Mozilla using it as a dairy cow, and starving it on top of that.


  > What I don't like is Mozilla using it as a dairy cow, and starving it on top of that.
I don't get the analogy. What exactly is Mozilla doing that you don't like?


They’re implying that Mozilla are milking Firefox for all it’s worth. Exploiting it and at the same time not treating it very well.


Correct. I feel Mozilla is starving Firefox and using the money somewhat frivolously.

Mozilla have many interesting projects and several of them may be good, but none of them have had such impact or has such potential for the future as Firefox.

Edit: and I'm willing to pay $10 - $50 a month to someone who will create and maintain a patched version of the latest Firefox that fixes the worst problems like not being able to hide the standard tabs (in addition to any sponsored search deals they may get).

I suspect I'm not alone: for many(most?) of us our browser is one of our most important tools, the other being and IDE, an editor or some graphics tool.

Edit 2: Paid Chromium based doesn't count for me. A major point is to counter the Chrome monoculture.


Well, that's a really bad reason to use chrome or some other chromium based browser.

But if you are willing to fork Firefox, that's a very good reason. I just don't think any fork will be as long as the majority of users are on a chromium based browser, those are toxic to the ecosystem.


> Well, that's a really bad reason to use chrome or some other chromium based browser.

Who said I do, I often have multiple months between every time I use a Chrome or Chrome based browser.

Edit, I use a multipronged approach:

- I use only Firefox - except once in a blue moon to verify if something is an actual Firefox bug or a general bug.

- and develop in Firefox. Bonus: Without testing in any other browsers most weeks I can count on one finger the times I have introduced cross browser defects

- I raise awareness that Mozilla is extracting money from Firefox, not funding it.

- I raise awareness about how Google is pushing to kill competition in the browser markets (besides here on HN and contacting authorities myself I have also urged a grumpy colleague today to contact relevant competition authorities)

- I rise awareness about the likely outcome of a Chrome monoculture: mostly that ad blocking will disappear, the web as a platform will stagnate and we will have to live with more nasty restrictions.


Firefox opens new tabs saying “We care about your privacy, look, LOOK!” every time you start, sometimes two of those tabs (release notes + privacy). I wish I could just deactivate those built-in ads.


Recently, they had this annoying modal ad (with no "close when clicking outside" feature) telling me that I can make my browser colorful. I care about web-health and containers is far too much of a must-have feature to make me switch, but I’m finally starting to get annoyed.

Without containers, I’d probably bite the bullet and start using some chrome fork that doesn’t show me useless ads.


Recently they also changed the urlbar to do a search when you type anything... even localhost. INFURIATING. To stay polite. Messed with about:config but did not find a way to disable that crap.

So Chrome and FF are in the same boat: "UX" "designer" taking non-sensical decisions for the whimsical greater good.


I do not have this issue, but I have a very customized about settings.

Here is the relevant about:config settings I have these changed for the URL bar:

  browser.urlbar.suggest.searches false
  browser.urlbar.searchSuggestionsChoice false 
  browser.urlbar.showSearchSuggestionsFirst false
Also for your urlbar you want to change it so it always shows the scheme and every part of the URL.

  browser.urlbar.trimURLs false
Stop Firefox trying to help with incomplete urls and loading the wrong site:

  browser.fixup.alternate.enabled false
Setting the above about:config entries should stop URLS you type being sent to a search engine and also stop some other surprises in the URL bar.


I do not have this change.


Me neither. And this is the first time I've heard about it; such changes often raise a lot more havoc on the social media outlets.

I presume grandparent has either hit a bug, has some unusual/untested about:config combination, or, most likely, an addon that broken.


The Settings have several checkboxes to switch of (or on):

* Tips and News from Mozilla and Firefox * Recommendations while surfing * Recommend addons while surfing

I'm not entirely sure what you saw that made you angry, but I'm pretty sure you can switch them off rather than abandon the entire browser.

That goes for a lot of Firefox hate, I find: quite often people are ranting online about some new or removed feature, which they can dis- or enable easily in the settings. Or -a tad harder- in about:config. Or even a tad harder, with an addon.

That makes me think those ranters don't really want their problem solved, but just want to vent some anger about X changing something that they are emotionally not ready to see changed.


> Tips and News from Mozilla and Firefox

Was switched off.

All the others (according to the linked information) are irrelevant for intrusive pop-up ads. By all means, please explain to me which of those settings do something different from what they are supposed to do, or tell me about an about:config settings that tells Firefox to never show me pop-up ads.

What made me angry: https://i.imgur.com/s9hC23U.png which blocks the interaction with any part of Firefox until I click "Not Now"


Is this not a one-time thing that's shown after every update? Or is it recurring?


It was one time for me. Although I do think what they advertised there and the way they advertised it was nonsensical and disrespectful to the user.


Sounds like a bug then. Is there a bug-report for this?


Firefox for iOS has NO about:config feature and a paltry amount of settings under Settings submenu.


Weirdest criticism I've heard. Either someone know it's all Safari in a skin (due to Apple's restriction), or someone didn't heard of this limitation.


How much does all iOS browsers being built on top of Safari actually limit what one can do?

Obviously the actual page rendering and JavaScript executing would be Safari. But 99% of the time when I hear people advocating for browser X over browser Y it is not because X has better rendering or a better JS engine. It is because of higher level things, like containers (Firefox over Chrome) or better profile handling (Chrome over Firefox) or better spell checking (everything over Firefox).

Does having to build on top of Safari on iOS also constrain those higher level things?


> Does having to build on top of Safari on iOS also constrain those higher level things?

Yes!

> containers (Firefox over Chrome)

No, they can't install add-ons (https://support.mozilla.org/en-US/kb/add-ons-firefox-ios). Even if they could built this into the browser, they can't (see how the profiles are handled).

> better profile handling (Chrome over Firefox)

No, Chrome does not support this feature on iOS (https://support.google.com/chrome/answer/2364824?hl=en&co=GE...). iOS' WebKit limits this to only one permanent profile and one ephemeral profile. You can technically have multiple profiles in the iPad - that is, if you're a school, and that's really more of a OS-wide user account thing. (G did a hack over this by rewriting that profile when switching profiles - but it's not truly a multitask thing, and violates Apple Developer Guidelines).

> better spell checking (everything over Firefox)

No, and for more sensible reasons. Users do expect consistent spell checking because they use the keyboard to do that. Regardless, if you want to implement spell checking in-browser, that's impossible. If you want to do it outside of Safari, go ahead - it's just not integrated to those Apple things (more of a dedicated Grammarly interface rather than desktop spell checking).


Yes, I got that one too and was extremely perplexed that it didn’t close when I clicked outside it and I had to hunt the tiny x instead.


I found the colored browser cool, but the infuriating part is that Firefox is deliberately only offering features for a limited time. And if I didn't want to pick a color, not being able to click outside to close would be annoying.


Unlike many here, I don’t mind FF giving their browser more mass-appeal (here in Germany it’s still a major browser), so adding that color feature is fine. I just don’t need intrusive advertisement about it. It’s even worse for me, because I use FF on multiple machines and with several profiles, so I saw the ad over 10 times in total.


Baffling. You replied to my statement that you can easily change those with a rebuttal that you don't like the defaults


Can I easily disable the post-update splash page telling me how great the latest Firefox is (or something like that)? I use developer edition to test Firefox compatibility of my webapps, I seem to get that a lot, and I really don’t care about that marketing.


On the contrary, there is no option that says “Stop bugging the user with ads for Mozilla.” They keep coming back with another form of showing up (button in the top bar, label on the new tab experience, new tab experience itself, Mozilla login…) and you have to disable it again. It’s like spam when they use several Mailchimp accounts, reupload your email and tell you “but it’s easy to unsubscribe!”


Does Firefox actually get hate? I don't think I've seen people actually make digs at FF.

For me it would be Edge that gets the most laughs but I find is a better performing browser, at least in terms of UI than FF or Chrome. Side-bar tasks, grouping, integrated screen-shot, etc.


Mozilla management gets hate. They have had a string of controversies, policy upheavals that have bad optics and layoffs etc. They are often on HN for all the bad reasons.


Maybe not Firefox itself but Mozilla’s “We need more than deplatforming” blog post burnt a lot of good will towards them. I think that kind of censorial ethos is at odds with the majority of people who would specifically choose firefox over chrome.


Firefox is significantly less secure than Chrome, though it may be more private.


We need facts tho prove this



Interesting article but the issue is that it's mashing together the Chromium Browser and Google Chrome. I'm sure that Chromium itself is safe but the prvacy concers arise from the "Google bits" rather that the browser engine itself.


Right, which is why the article mentions that it is focusing on security and not privacy.

With a bit of effort, you can get a similar level of privacy to Firefox in Chrom[e/ium].


>With a bit of effort, you can get a similar level of privacy to Firefox in Chrom[e/ium].

What specific effort?


Some of the statements in this article are just not true. The post seems to get frequent edits and always looks new/recent when I visit it, though I’ve seen and rebutted it several times over the last years. Yet, apparently nobody made the effort to verify the claims with more recent analyses.

A real shame.


Could you point out some untrue statements? The claims seem to be well-sourced.


Thank you


Please elaborate



Chrome used to have an option of a blank page, but it was removed "for user convenience"


Fortunately they made an extension for it: https://chrome.google.com/webstore/detail/empty-new-tab-page...


Until it gets removed "for user convenience"


Chrome disabled the extension for me because of "chrome store policy". I was graciously allowed to enable the extension again...


citation needed... afaik it's always been available via an extension and never via a setting


TIL


[flagged]


Or how will my innocent friends learn of my sordid browsing history? -_-


This seems oddly specific. Show me where the insightful but repetitive, moderately funny British man hurt you.


> Dhone took away a $1,000 bug bounty reward

Damn that is an offensively small amount - less than the cost of a Google engineer for a days work.


It wasn't a particularly likely exploitation route... The user had to already be double-clicking files they'd downloaded from a malicious webpage. At that point, it might as well have been an .exe file.

And after all that, all it can do is run a search query. It can't leak all your Gmail emails or exploit the local machine.


> And after all that, all it can do is run a search query. It can't leak all your Gmail emails or exploit the local machine.

Doesn't that contradict the following?

> “However, because the IPC channel was exposed to JS directly in New Tab page, the XSS in Chrome’s NTP can be treated as the equivalent of renderer process RCE.”


Yes, but that IPC channel to the browser can only do anything useful if you find another exploit in the browser.

And the New Tab Page doesn't even have permissions to do much via that IPC channel, because its origin isn't equal to anything interesting.


The xss is stored via csrf. So an exploitation scenario would be that you visit a malicious page and then click in the search box on a new tab.


I also hate not having option to make my new tab page empty but thinking about the time I have spent for people I know to make their browser homepage cleared ... I won't object it being managed by the browser companies... if you know what I mean... mendokusai...


Google VRP is giving the wrong incentives here, as such a small (insulting?) reward will surely orient some researchers to exploit market rather than responsible disclosure.

Google, shame on you.


Do you have enough domain knowledge to be judging the incentives ? What do you think would be a fair amount ?


Q: Do you have enough domain knowledge to be judging the incentives ?

Well... I don't know. Does anyone have to be a domain expert to say that security reporting that affects tens or hundreds of million of people should be compensated better than 1k USD?

I dislike a bit the "justified" argument, as very often it dismisses important weak signal warnings. Our work in Security is often about being sensitive and not dismissal. But here you go:

I'm infosec since 1987 (34 years) and never left it, so I'll let you decide ;-) even if i'm a dinosaur in Internet times ;-)

Q: What do you think would be a fair amount ?

IMHO, the fair amount is definitely in the tens of thousands.

But we could attempt a quantified approach, always debatable (Risk = Likelihood * Consequence), eg. Likelihood based on fishing campaign success per country or global, and then mean / average cost of theft when leveraging the full exploit chain (IPC included), i.e. cookies -> auth -> leveraged identity theft impact. And then give percentage of cost as an bounty-based "insurance" mechanism. Not easy but attempt could be done. Surely that would result in way higher compensation.


The Likelihood of this bug being exploited is damn low. It requires the download of an html file, and then the target would have to double click it. On top of that the Consequences of it are not that serious.so you provided you answer as to why the bug only got $1000


I'd be interested to know what the market would have paid for this bug. I don't really see why it would be useful to anyone but I am far from an expert.


Feels like it could make a decent "tracker" of sorts, phoning home a ping on every new tab open. More useful if it works on mobile chrome.




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

Search: