Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Pile Of Garbage
May 28, 2007



Grog posted:

Not sure if I should just ask in here or make a new thread, but I'll ask anyway. Let me know otherwise.

I downloaded a program. Being generally paranoid, I did an on-demand scan (like usual) on the executable/self-extracting archive and got a slightly inconclusive hit (seems like potentially a false positive, from what little information I found). I never manually ran the executable. Uploaded it to VirusTotal and 15/55 came back with a result but again, the results seemed fairly generic and inconclusive. I deleted the program and have run several full scans on my system with MSE, TDSSKiller, Rkill, Comodo Cleaner Essentials/KillSwitch/Autorun Analyzer, Malwarebytes, and SuperAntiSpyware. Running KillSwitch, there are a few drivers in the "Services" with nonsense names that I can't find any real information on (qrhwas, tcoifh, uotote), but I have no idea if they're related to the various scanners or what. Other than that, there are various "unknowns" from my Bluetooth and Creative audio drivers, and some mshtml.dll entries from Autorun Analyzer (about, javascript, mailto, res, vbscript), but they're signed Microsoft and don't throw any flags in the other scanners. Nothing else unknown or suspicious shows up in any of the other scanners.

There aren't currently any new problems that I've noticed with the system. Is there anything else I should do to try to make sure nothing's wrong? Should I still be at all worried about this, or should I just assume it was a false positive and the system's fine? I tend to be overly paranoid and just wipe everything whenever a suspicious file pops up, which I think is probably overdoing it and overly time-consuming.

What did the AV engines on VirusTotal report it as? Usually you can find an accompanying KB article from the relevant vendor which lists signs of infection that you can check for. Of course that's assuming that it's fairly benign.

Adbot
ADBOT LOVES YOU

Pile Of Garbage
May 28, 2007



Khablam posted:

So with SSL fuckery (thanks Dell) and manufacturers doing MITM attacks on their own customers, bad AVs self-signing your requests (breaking EV) should we talk about SSL security?

The GRC page probably best outlines the basics and offers a way at testing your results: https://www.grc.com/fingerprints.htm
The perspectives project is available for firefox which seeks to do the same on the fly - http://perspectives-project.org/

There also seems to be a few tools to check your existing stores - http://www.wilderssecurity.com/threads/rcc-check-your-systems-trusted-root-certificate-store.373819/

Does anyone know of a better means of verifying a systems SSL integrity?

That GRC page is absolutely garbage. SSL Labs is pretty much the go-to for validating the strength of a web-servers TLS implementation/configuration.

Edit: it appears that you are actually talking about PKI and validation of client-side trusted root/intermediate CA certificate stores, not SSL/TLS. Just keep your OS, web browsers and any other applications which maintain their own trusted CA certificate stores up-to-date (e.g. JRE). If you've bought a new machine then it really depends whether or not the OEM is a massive rear end in a top hat so format and reinstall the OS I guess.

Pile Of Garbage fucked around with this message at 14:20 on Nov 25, 2015

Pile Of Garbage
May 28, 2007



TwystNeko posted:

Similar to Goodpancakes, I've had several login alerts with google.

In fact, it's been a hell of a weekend. :(:

Saturday morning, at 1:30, I get an email from paypal that my order for a brand new iphone has been completed, to the tune of $1000 CAD. I didn't order this. I change my passwords, go through all the account settings, and file a dispute. It gets cancelled by the rear end in a top hat who ordered it, so I spend 90 minutes on the phone with paypal support.

I check my ebay - turns out, that's how they got in. Since my paypal was linked, that's all it took. That's all been changed and disabled too.

And tonight, I've had 3 alerts from google that someone has my password - Even after I've changed my password. Changed it every time.

I've done a malware scan, I've done a virus scan. I've checked all running processes to ensure I don't have anything I don't recognize running. I run uBlock Origin, and checked all my extensions. I seem to have a clean system, but who knows?

So at this point, I think I need to start using autogenerated passwords via a manager of some sort. Or 2FA. Or both.

Things I want out of it: Android app / browser integration (Currently chrome, but I'll gladly switch) / 2FA support. I don't mind paying for a subscription. I keep seeing LastPass recommended everywhere but here - 1Password seems to be $65 USD currently (which is like $90 CAD), and KeePass confuses me a bit.

Help?

Check the apps which have been authorised to access your Google account here: https://security.google.com/settings/security/permissions. If there are any that you don't recognise then revoke them (Or just revoke all of them). Enable 2FA on your Google account using the Google Authenticator app (https://play.google.com/store/apps/details?id=com.google.android.apps.authenticator2&hl=en) as your primary and SMS/landline as backup. Then enable 2FA on all other services which support it (If they support OAuth2 then you can hook them in with the Google Authenticator app). Once you've done all that you can start looking for a password manager (KeePass seems to be popular, I can't comment because I just use Password Safe on an encrypted volume like some sort of old).

Pile Of Garbage
May 28, 2007



Hammer. Seriously, if you have data worth recovering from a physically damaged HDD by a nefarious third-party then you wouldn't be posting here.

Unless you want to sell the HDDs?

Pile Of Garbage
May 28, 2007



OSI bean dip posted:

The more destructive the better.

Sell them to Colin Furze for thermite cannon target practice!

Pile Of Garbage
May 28, 2007



Just smash it with a hammer. Unless you have to conform with a data disposal standard just hammer it.

Pile Of Garbage
May 28, 2007



Segmentation Fault posted:

I was under the impression it did some other things too. Is uBlock Origin just a superset of Disconnect then?

I don't know whether there's any overlap between the uBlock and Disconnect filters but you can enable a whole bunch of different third-party filters, including Disconnect, in the uBlock Origin add-on settings:

Pile Of Garbage
May 28, 2007



Might want to add a bit about restricting administrative access (HTTP, SSH, etc.) on the device so that it can only be administered via your internal network. Heaps of garbage devices have that enabled by default and it's how they get owned.

Pile Of Garbage
May 28, 2007



Squeegy posted:

This thread is neat and I've enjoyed reading it. It's also gotten me to tighten up some stuff. I have a few questions; OpenDNS seems to be widely recommended, but I've been using something called Simple DNSCrypt. Have you ever heard of it and is it worth using over OpenDNS? It encrypts your DNS traffic, which seems like it would help prevent MITM attacks, and hasn't had any noticeable downsides for me other than occasionally changing the server when things stop loading.

Encrypting the connection from your client to the DNS server may make MITM more difficult however most malware simply just tampers with your DNS client settings or edits your host file.

Edit: to actually provide some advice, if you have a firewall then block all outbound udp/53 except for DNS servers that you trust.

Pile Of Garbage fucked around with this message at 11:50 on Sep 2, 2016

Pile Of Garbage
May 28, 2007



Mo_Steel posted:

Relevant side-question from the discussion on the last page, anyone got a good breakdown of HTML5 vs. Flash in terms of security and vulnerabilities? Is it sandboxed better to prevent egregious poo poo, less long-standing known security flaws, smaller current userbase, etc.?

They're really two different things entirely.

HTML is just a markup language which your browser parses and feeds into its layout engine to render the page. Videos are simply embedded using the <video> element which contains a URL to the video file which the browser downloads and plays back. The browser doesn't actually "execute" anything (Ignoring JS).

On the other hand, Flash objects (SWFs) are essentially compiled applications which are executed by the Flash Player plug-in. This is the main reason why Flash is inherently dangerous, you're executing untrusted code on your machine.

Pile Of Garbage
May 28, 2007



OSI bean dip posted:

I would never ever trust a computer. If you can find a way to flatten them, you should do so.

ftfy

Pile Of Garbage
May 28, 2007



uPen posted:

You're not going to run into something that can turn a USB drive into an infection source on 99.9% of computers. Copy any documents, pictures, music etc you want off the pc, format the drive and reinstall windows.

Yeah this is bunk. Any file system mounted RW on an infected machine should be considered compromised.

Pile Of Garbage
May 28, 2007



African AIDS cum posted:

My PC was recently hacked or infected with a trojan or something, not sure as no scanner picked anything up, but someone was able to get into my email/amazon/banking etc, even bypassing 2 factor authentication. Also noticed a ton of bandwidth being used by tcpsvcs.exe to some random brazillian IP address. I am very careful about what I allow to run, so I really have no idea what happened.

I did a clean install of windows 7 on a new SSD. I am wondering if using Linux in a VM for all email, banking etc. would be safe, since dual booting would be a pain. Or maybe I should just go back to Mac.

Doing your banking, etc. in a VM will only be "safer" if you already assume and expect your PC to be compromised. If your PC has been compromised then they would have probably owned your accounts anyway. In addition things like phishing attacks rely on exploiting the user more than the computer so if you aren't being careful you can be owned regardless.

Upgrade to Windows 10, keep your OS and software (Browsers, etc.) up-to-date, enable click-to-play for all browser plug-ins (Flash, Java, etc.) or if you don't need them uninstall completely, get an ad-block extension for your browser (For FF/Chrome: uBlock Origin), get a password manager and set different passwords for all your accounts online, enable app-based 2FA for everything and be more careful online.

Pile Of Garbage
May 28, 2007



Cup Runneth Over posted:

What the absolute gently caress, Adobe? I can't save documents in Adobe Acrobat Reader DC without triggering EMET's Caller mitigation and crashing the program. How do you gently caress that up??

More often than not it's the AV that's hosed up but we are talking about Adobe here soooo

Pile Of Garbage
May 28, 2007



Cup Runneth Over posted:

EMET isn't an AV.

Lol yes you are correct. For some reason I read EMET as ESET...

Pile Of Garbage
May 28, 2007



Subjunctive posted:

I would hope that it wipes all unallocated space.

This. Also depends on how you're wiping it (Pattern, number of passes, etc.)

Pile Of Garbage
May 28, 2007



Oysters Autobio posted:

Is this a good thread to ask about VPNs?

I'm currently revamping my entire home networking setup (previously I was lazy and used a slow router/modem combo given by my ISP), and came across certain routers that have built-in VPN capabilities.

Is it worth it to consider a router with a built-in VPN, or is there added security risks with using the same device for both?

The security risks will depend on the VPN implementation in software and how you configure it. IMO you should only consider it if you have an actual requirement for it.

RFC2324 posted:

don't you pretty much have to run the vpn on the router unless you want to expose an internal system?

I'd trust my router to have fewer vulnerabilities than an exposed system.

Never trust consumer networking equipment to not be riddled with vulnerabilities.

Pile Of Garbage
May 28, 2007



Blowdryer posted:

1. Enhanced Mitigation Experience Toolkit (EMET)
- I checked the link and I see "The new end of life date is July 31, 2018. There are no plans to offer support or security patching for EMET after July 31, 2018. For improved security, we recommend that customers migrate to the latest version of Windows 10."
I have windows 10, does this mean I don't need EMET?

Correct. The protection provided by EMET was rolled into Windows Defender.

Blowdryer posted:

2. At least for Windows' built-in firewall, configure it so it denies all inbound traffic regardless of what network you are on and only enable inbound ports if absolutely necessary.
- Would this mean opening windows firewall and literally setting all inbound traffic to none? Will my internet still work? Are there specific rules I should use? I assume I should unblock like steam or utorrent or things like that, but I'm just looking for a little more clarification on how exactly I go about accomplishing this.

Unless you're going to be using a 4G dongle or have a weird setup your PC will be behind NAT and not directly exposed to the internet. Modifying the Windows Firewall configuration will only really affect communication with other devices on your LAN. IMO just leave it as-is and make sure you will be fine, if you're super worried just set the network location to "Public".

Edit: applications like Steam and most Windows torrent clients are smart enough that they'll automatically add inbound rules to Windows Firewall when you install them. However, and this is more specific to torrent clients, if you've disabled UPnP on your modem/router (And you should) then you'll need to configure static NAT/port forwarding to allow inbound connections. Also uTorrent is terrible, use Deluge instead.

Blowdryer posted:

3. Make sure that admin access is disabled from the Internet
- How exactly does this work? It sounds like "don't have an admin account which has the ability to connect to the internet" but that doesn't really seem right so I'm probably understanding it. Can anyone explain?

This is referring to your router/modem and administrative access via HTTP/SSH/Telnet/etc. You need to check the configuration of your router/modem and make sure that it only allows administrative access from your LAN and does not allow access from the internet.

Blowdryer posted:

4. Should I get Malwarebytes Anti-Malware? Planning on using Windows Defender & Firewall.

No, don't bother with MBAM or any other AV really as they're all hot garbage.

Pile Of Garbage fucked around with this message at 11:00 on Jan 6, 2018

Pile Of Garbage
May 28, 2007



Rexxed posted:

Malwarebytes anti-malware had a bug in an update today that caused some kind of memory leak that would cause windows to crash. Apparently there's a new update that fixes it but if you had windows problems today and use it, that's why:
https://forums.malwarebytes.com/topic/220112-important-web-blocking-ram-usage/

cheese-cube posted:

No, don't bother with MBAM or any other AV really as they're all hot garbage.

Pile Of Garbage
May 28, 2007



IMO the most important aspect of using a password manager is that it encourages you to use unique passwords for each service that you interact with. If a service you use is doing something stupid like storing passwords in plain text it won't matter how complex your password is if their DB gets popped. However if your password is unique to that one service then it will prevent whomever from compromising your other accounts.

Realistically having 2FA enabled with an OTP app is more important than making sure your password is complex.

Pile Of Garbage
May 28, 2007



yoloer420 posted:

Advertisers can target you based on a variety of mechanisms, this is true. Google won't give that information or your details to the advertisers though. That would be against Google's interests because then advertisers could target you directly, without continuing to pay Google.

You can generally trust Google to protect their income source, which thankfully means not handing out people's personal info.

Yeah but if they ever did find a way to make more money by handing out people's personal info then they'd probably do it.

Pile Of Garbage
May 28, 2007



OSU_Matthew posted:

5. Pi hole is easy and effective to set up, and using Quad-9 as your DNS can help protect you from malicious resolutions. I posted a few pages ago about this, and it was just brought up again. Not a panacea, but it helps

Quad-9 had issues with geolocation for services like Office 365, not sure if they've since rectified them.

Pile Of Garbage
May 28, 2007



I didn't see it mentioned in the Cloudflare blog post but 20k prefixes were affected by the issue. Good stuff :lol:

https://twitter.com/bgpmon/status/1143149817473847296

Edit: also here's a NANOG post that points out that the issue could have been avoided if poo poo was configured correctly: https://mailman.nanog.org/pipermail/nanog/2019-June/101589.html

Pile Of Garbage fucked around with this message at 04:31 on Jun 26, 2019

Pile Of Garbage
May 28, 2007



olives black posted:

- Configure unprivileged user account for general use
- Configure UAC to require admin password for all actions requiring admin privileges

This adds little in the way of security and will drive you insane. Just leave the UAC config as default.

olives black posted:

- Configure firewall to block all inbound traffic that is not pointed at ports 22 and 5001 (for sshd and Universal Media Server, respectively) from unapproved IP addresses (my PS4 and a few other things maybe) or part of an established connection

The default Windows Firewall profiles block almost all inbound protocols quite adequately. You'll be fine unless you're port-forwarding inbound to your PC on your router and/or your LAN is packed full of dodgy devices (If it's the latter then the best approach is to put your PC on a separate VLAN and permit/deny traffic on your router).

olives black posted:

- Replace default browser with NoScript-hardened Firefox

I've never seen the point of NoScript tbh. Just get uBlock Origin and don't install Flash or Java.

olives black posted:

- Install whatever Microsoft's official antivirus is atm assuming it doesn't already get set up with the OS installer

No action required really, Windows Defender is enabled by default in Win10 and the default settings are OK. There are some extra hardening settings you can enable like Controlled Folder Access which is neat.

olives black posted:

- Disable Cortana and all of the other search bullshit

This is fine however please don't run one of those PowerShell scripts which nukes all the AppX stuff as those honestly do more harm than god. IMO best approach is to uninstall what you can and just not use the rest. It's also worth going through every section in the Windows Settings (New Control Panel) to disable all the dumb telemetry stuff.

olives black posted:

- Change DNS service to OpenDNS or something else (more research required)

OpenDNS got bought by Cisco and rebranded as Cisco Umbrella. Generally for a home setup the recommended approach is to use anything other than your ISP for DNS resolution. The main options out there are:

  • Cloudflare DNS: 1.1.1.1 & 1.0.0.1
  • Google Public DNS: 8.8.8.8 and 8.8.4.4
  • IBM Quad9: 9.9.9.9 & 149.112.112.112
  • Cisco Umbrella: 208.67.222.222 & 208.67.220.220

I personally use Cloudflare as it's the fastest and Google as a backup. Quad9 and Umbrella apparently offer enhanced security (Blocking malicious domains, etc.) however I've personally seen GeoIP issues with Quad9 and Umbrella is just kinda bad.

You might also want to look into DNS over HTTPS (DoH). Cloudflare's resolver supports it and you can enable DoH in the latest Firefox release however there's not exactly any support for it at the OS level in Win10.

Pile Of Garbage
May 28, 2007



olives black posted:

Thank you for the follow-up!

I disagree with the first point (it gives me peace of mind knowing that I can't just click OK out of frustration on something dangerous and I'm used to using sudo on Linux anyway), but the rest has given me good stuff to consider.

That's the thing though: are you confident that when presented with a UAC elevation prompt, expected or otherwise, you can adequately determine whether or not the process which spawned it can be trusted (Remembering that whilst the UAC prompt is active you cannot interact with the OS and the prompt itself only includes at most the process name, PID and path)?

This is essentially one of those situations where you're sacrificing usability in the name of security without actually increasing security.

Also your analogy isn't really accurate as sudo is manual elevation and more akin to "Run as Administrator" on Windows.

Pile Of Garbage
May 28, 2007



olives black posted:

Point taken regarding the sudo analogy. However, it's less about knowing whether or not I can trust the process as it is about forcing myself to slow down and consider what's happening.

Worst case scenario is that I say "idk wtf", cancel out of it and have to investigate what's going on afterwards.

I feel like you maybe don't use Windows on the reg and/or haven't used it in some time because honestly in Windows 10 UAC elevation prompts are very rare. I use Win10 all the time at work and at home and only ever see UAC prompts when I'm installing stuff which is when I expect them. Outside of that I never see them but if I did I'd know something suss was going on because well it just doesn't happen.

That aside I guess it's up to you but at the end of the day if you were to get infected with something it'd probably be using some UAC bypass exploit anyway :lol:

Pile Of Garbage
May 28, 2007



olives black posted:

Holy moly, I forgot what a clusterfuck Windows Firewall is. Is there a way to flush all of the default crap and make a configuration similar to this where inbound traffic is limited to established connections and whitelisted ports/IP addresses only?

If you run Disable-NetFirewallRule with no parameters in PowerShell it will just disable all the rules in all three profiles (Public, Private and Domain). That aside I'd like to reiterate something I mentioned earlier:

Pile Of Garbage posted:

The default Windows Firewall profiles block almost all inbound protocols quite adequately. You'll be fine unless you're port-forwarding inbound to your PC on your router and/or your LAN is packed full of dodgy devices (If it's the latter then the best approach is to put your PC on a separate VLAN and permit/deny traffic on your router).

Unless your PC is directly exposed to the internet I really don't think you'll get any real benefit by mucking around with Windows Firewall policies.

Pile Of Garbage
May 28, 2007



If there is stuff that you must simply keep on the PC then remove the HDD, swap in a fresh one, reinstall Windows and then mount the old one to a Linux or something to pull files off.

Also make sure you go through your mum's accounts and remove any app access (That poo poo where an app gets issued a token for perm access to the account). My grandma's FB account got compromised a while ago through one of those. Changing passwords and enabling MFA didn't do poo poo because the mechanism popping the account had access via the token.

Pile Of Garbage
May 28, 2007



The important thing here being to not connect the potentially infected drive to the newly reinstalled system because then you'll potentially just undo everything. Attaching the drive to a Linux system is usually good as it's very unlikely that whatever is on it would be able to pop a Linux system.

Pile Of Garbage
May 28, 2007



Carbon dioxide posted:

Does the OP's section about antivirus products apply to ClamAV on Linux?

Is it at all useful to install ClamAV on a Linux system?

Yes the OP applies to ClamAV because ClamAV is an antivirus product.

I personally don't see any benefit to installing ClamAV on Linux systems beyond box-ticking asinine compliance reports.

Pile Of Garbage
May 28, 2007



Is it only going to be used with Windows? If so then BitLocker is perfectly fine.

Pile Of Garbage
May 28, 2007



uguu posted:

Isn't bitlocker only for professional and enterprise on vista? I have home premium. I know there are lots of ways around the problem, but I'd like to keep the drive as is, frozen in time, if possible.

Wait why are we talking about Vista?

Assuming a mistype IIRC the TPM-backed version is only available on Pro or Enterprise but but the password-based "BitLocker To Go" is available on all versions.

Pile Of Garbage
May 28, 2007



I'm certain if it's a boot volume you can't use BitLocker To Go. And if it's not on Win10 I think it uses some algo that's not backwards compatible.

Maybe if you actually told us what you are trying to achieve then we can give good recommendations?

Pile Of Garbage
May 28, 2007



tuyop posted:

The upgrade to 10 pro is the proest recommendation, thanks. I didn't know you could do bitlocker without some kind of TPM so that's very helpful!

For reference Microsoft call it BitLocker To Go.

Quaint Quail Quilt posted:

I've recently done encryption with pro and the only time I know it's on is if you reinstall windows or update the bios you have to type your 100 digit key in.

Pulling it up on your phone and knowing how to use a numpad without looking helps greatly.

That's regular BitLocker using TPM.

Pile Of Garbage
May 28, 2007



Oysters Autobio posted:

So I've been using Dashlane for almost a year now but am not too happy with the functionality (the autofill on it isn't really good so it rarely actually pops up to automatically prompt password changes or new accounts). I only really subscribed because they offered a VPN alongside the subscription so I thought why not. Also found out that the VPN they contract out to is apparently sketchy in terms of selling your log data.

Among the Goon-approved password managers, are there any that can hit all or most of these features:

1) Built in VPN? I don't use a VPN that often where I feel its worth it to subscribe to one but when I do (like using public wifi for example) I like to be able to throw it on and also not hit any data limits like the free VPNs offered out there.

2) Easy import of Dashlane passwords and Chrome passwords

3) Password monitoring that searches leaked db's

4) Built-in password generator, non-password notes that can be added

5) Good UI and integration with Chrome (i.e. good autofill and prompts when it detects new accounts or pw changes)

6) Can be used and syncs across multiple devices.

Also two optional,

- it has a decent shared plan for two people.
- can detect/change/autofill passwords in non-Chrome apps like Steam (Steam keeps constantly forgetting my password) or android apps

IMO you'd be better off looking for password manager and VPN services separately. They're solutions to two entirely different problems and as such there's a lot of outfits that do one or the other well but almost none that do both well.

Also password monitoring isn't really super important if you're using a password manager properly because then you will have unique passwords for every service which mitigates the risk of a service being hacked. If you really want it then do as hooah recommended and sign-up for Have I Been Pwned notifications.

This aside I don't really have any recommendations sorry. I run my own VPN with dedicated hardware and only use Password Safe on my home PC (Never really found the need to sync my creds).

Pile Of Garbage
May 28, 2007



Midjack posted:

Keyfile in a safe deposit box if you have one, or encrypted and given to your attorney or a friend or family member whom you would trust not to lose or try to crack it is an enhancement to this scheme, though depending on your circumstances you may not have any of those or wish to expend the resources to get them.

I think it's worth repeating that whenever you go down the road of securing your things you should have a good think about what your threat model actually is. This includes evaluating yourself as a target.

I'd be willing to bet that the large majority of people here on SA, myself included, are effectively nobodies with little in the way of valuable assets. If you fall into that category, and you probably do, then you're a low-value target and only need to worry about the same poo poo as everyone else: scams, malware and phishing.

If you're in this category then leaving poo poo like encryption keys or MFA recovery codes in a safe deposit box or with an attorney offers no more protection than leaving them on a piece of paper or a HDD at home.

Further to considering yourself as a target you need to consider what possible threats you might face because there are many that you as an individual can't do poo poo to mitigate. An oft used example is if Mossad wanted your AES 512-bit encrypted data. In that case they'd simply kidnap and torture you to get the key! That of course isn't a threat any of us would face but it goes to show how many security measures kind of fall apart if you're just an individual without corporate/government backing.

To summarise, just go with what is sensible and easy. Unless you're Alexei Navalny enabling MFA on everything, using unique passwords for every single service, using a password manager, running ad-block, installing software/OS updates on your phone+PC and not clicking links in e-mails will protect you from 99% of potential threats. Anything more is probably unnecessary.

Pile Of Garbage
May 28, 2007



That's a good point, I am a massive dingus!

Pile Of Garbage
May 28, 2007



Wait a minute this is all dumb. If the risk is "a fire obliterates your house and all your belongings" then your lock-box key is probably toast as well. Just get a fire-proof safe or put your key on some cloud service that's separated from your other stuff.

Khablam posted:

The best way to have a keyfile is obfuscation. Don't have "my keyfile.keyx" and decide you have to hide it, just use some commonly available file, but I repeat myself.

So you're advocating for security through obscurity? You do know forensic file system tools exist right? They can look through a whole bunch of files and immediately ID anything that's out of place (NTFS Alternate Data Streams, weird EXIF data, file data structure not matching header, other sneaky poo poo I don't know about plus most of the common steganography techniques).

If you really wanted to hide the file in such a fashion you'd have to be real fuckin sneaky about it, so much so that you'd end up having to record where and how you've hidden it because there's no guarantee that you'll remember whatever arcane process you used.

Pile Of Garbage
May 28, 2007



Cup Runneth Over posted:

Huh??? Under what circumstances? Either you're home at the time and A) die in the fire, or B) evacuate with basic belongings like your keys and wallet and stuff, or you're not at home and you have your keys and wallet and stuff with you. In both cases you're either dead or have your keys, and why wouldn't your lockbox key be among them?

Complete destruction of all the poo poo was the premise Midjack posted up-thread:

Midjack posted:

Zorak of Michigan posted:

If my house ever burns down and I leave without my phone, I'm pretty well screwed.

Key file held off premises addresses house burning down with phone in it.

I agree it's probably unlikely but disaster planning is about assessing likelihood, impact and mitigation together.

Khablam posted:

A keyfile is just a second-factor that stops someone accessing your vault if they can a) get a copy and b)get your password. Nothing about using a holiday snap of a pier makes it any less secure. If they can ID and lift your keyfile, it also doesn't matter what it is.
Which means it doesn't matter what your keyfile is, only that you know what it is.

Given this there's no reason to try to physically lock it away - as long as you know which of 20k photos you can restore from anywhere and use, you can... well use it. Same goes for any consistently hashable file you can get and use from anywhere.

I say this because for 99.99% of people, they're going to reduce their ability to use their own password vault far before they reduce their security.

Correct me if I'm wrong but are you suggesting that the keyfile be stored adjacent to the vault? If that's the case then you're effectively nullifying any benefits that a keyfile would provide. The whole point of a second-factor is that an attacker doesn't have it.

B-Nasty posted:

On Windows machines, a better approach is to lock down the key file, not obfuscate it.

On my machines, I set access to that keyfile to admins only, and set KeePass.exe to run elevated (UAC prompt). It's nice that it's a code-signed exe, because you get the blue UAC prompt with the author's name.

This accomplishes a number of things: malware.exe that I accidentally kicked off in user space can't read the key file, and process isolation in Windows means the memory used by KeePass (elevated process) is also protected from user-level process snooping.

A few issues here:

  • Restricting NTFS permissions doesn't mean much if someone can elevate to SYSTEM context.
  • Deliberately running KeePass as administrator provides no additional security and actually makes you more vulnerable. In the event that the EXE is compromised and you accidentally launch it as administrator, most likely through muscle memory of clicking through the UAC prompt, it will be in a context that it can compromise and root your system completely. Regarding the likelihood of your KeePass EXE being compromised, supply-chain attacks against popular software packages are far from uncommon so inadvertently installing a malicious update is a real possibility.
  • Most malware include privilege elevation exploits so relying on user space to constrain them is not guaranteed.
  • Process isolation is meaningless if malware can gain SYSTEM context.
  • There are also side-channel attacks, like simply reading the clipboard when you copy a password.

IMO a better approach is to store your key file on a separate volume encrypted with BitLocker (Passphrase AKA BitLocker To Go, not TPM) and to enable Ransomware Protection to control access to the volume/folder containing the key file (Assuming you're running Win10 which you should be). Furthermore you should ensure that KeePass is configured correctly, specifically the automatic lock and clipboard clearing options.

Edit: should probably add that the Ransomware Protection/Controlled Folder Access feature in Win10 isn't infallible so as always make sure to keep your system up-to-date.

Edit 2: if you're using Ransomware Protection/Controlled Folder Access and enable access to a location for say cmd.exe or powershell.exe you are basically screwing yourself. It's easy to shoot yourself in the foot so you gotta be careful.

Pile Of Garbage fucked around with this message at 20:27 on Mar 5, 2021

Adbot
ADBOT LOVES YOU

Pile Of Garbage
May 28, 2007



B-Nasty posted:

That's true, but kind of ridiculous. If malware can use an exploit to elevate or inject into a trusted process (these are rare), or you elevate it by UAC-clickthrough, you're hosed. Controlled folder access won't protect you at that point either.

Just tested it and it does block access by elevated processes to protected locations. Of course if you can elevate privilege you can modify the allow list. Also it only blocks writes and not reads so kinda pointless in retrospect (I'm constantly wrong lmao).

B-Nasty posted:

A risk, for sure, but consider that you're also trusting all your passwords to that application. At least with a local app like KeePass, you can choose to not install updates. With auto-updated browser extensions or online password services, you could be hit with this attack at any time.

I do agree that it would probably be more secure to store your keyfile on removable (or encrypted-by-default) media that you only mount when you want to open your PW DB, but that adds too much inconvenience for my liking.

Just to note I never advocated using any browser extensions or cloud password managers, rather just using a cloud storage service to store the key file, ideally one that isn't linked to any of your other accounts, the idea being to not store the keyfile adjacent to the vault itself because if you do that then the keyfile is no better than a password.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply