Login
You're viewing the functional.cafe public feed.
  • Oct 14, 2026, 9:25 AM

    Quite a few people have sent me the "Passkeys are a trap. Do not use them." post on Substack.

    The post recommends people abandon passkeys and stick to passwords + TOTP, citing what appears to be concerns with attestation as the primary reason.

    I’ve written a rebuttal as this is really dangerous advice and it's based on some incorrect technical claims.

    scotthelme.co.uk/passkeys-are-

    💬 11🔄 43⭐ 3

Replies

  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 2:23 PM

    @ScottHelme No one mentions that TOTP completely failed at doing what it was designed that it’s not phishing resistant

    💬 0🔄 0⭐ 1
  • Oct 14, 2026, 6:03 PM

    @plaka @ScottHelme Yup. TOTPs are a hack to effectively create unique passwords. It's hard to enforce unique passwords by policy, so instead we let users chose their own passwords and add a computer generated TOTP to it s.t. password stuffing won't work anymore. But phishing? Nope, TOTPs won't help you much there.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 2:45 PM

    @ScottHelme

    Here is IMH better argument against passkeys for personal use: hawksley.dev/blog/i-dont-like-

    To summarize: perfect fit for a corporate environment, but a poor fit for personal security.

    "To an individual, the greatest risks are instead permanent account lockout, automated account bans, and device loss. By using passkeys, you gain better security against man-in-the-middle attacks but face the higher probability scenario of losing access to your accounts."

    "If their [Apple and Google] automated systems decide one day to ban your account, you irreversibly lose access to all your passkeys used across all third-party accounts too."

    "I believe enterprise users have good reason to use passkeys, but the ecosystem isn’t mature enough yet for individuals."

    #passkeys

    💬 0🔄 2⭐ 9
  • Oct 14, 2026, 6:42 PM

    @maxpool

    Passkeys are an utterly unnecessary solution for corporate environments, because corporate environments already have PKCS#11, and have had it for 30 years.

    So no, passkeys are not good for corporate *or* personal use.

    @ScottHelme

    💬 0🔄 2⭐ 1
  • Oct 14, 2026, 2:52 PM

    @ScottHelme

    If anyone has the patience could they explain in plain simple English what the benefit of a passkey are over simple password, and 2fa where justified.

    I'm adverse to

    1. An overly vendor or consortium of vendor standards

    2. Someone else, especially an unknown unvetted 3rd party, managing all/most of my credentials

    💬 3🔄 0⭐ 0
  • Oct 14, 2026, 2:53 PM

    @rzeta0

    With a passkey you're not going to become the victim of a phishing attack, with a password and 2fa you can be.

    💬 1🔄 0⭐ 0
  • Oct 14, 2026, 4:34 PM

    @ScottHelme @rzeta0 But any reasonable password manager (that you say we should use for our passkeys) will check the domain before filling in the credentials?

    💬 0🔄 0⭐ 0
  • Oct 15, 2026, 1:02 PM

    @ScottHelme
    Isn't plain WebAuthn without any of the additional passkey features enough for this?

    Pseudonymous, not cross relatable between sites, and bound to domain and username it should be secure and private, and the only remaining annoyance is that losing the key is more likely than forgetting a password.

    @rzeta0

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 4:00 PM

    @rzeta0 @ScottHelme With a password, both you and the server you are logging into have a copy of your full password and possibly the seed for your 2FA. If that server is compromised, or you click on a phishing link and hand those credentials over to a hacker, your account can be compromised.

    With a passkey, you have half of your pass (private key), the server has the other half (public key). Only when they are together do they let you in and only on the website they were created on.

    💬 2🔄 0⭐ 0
  • Oct 14, 2026, 4:04 PM

    @rzeta0 @ScottHelme The device or password manager the passkey is created in will normally ask for biometrics as a second factor.

    My advice is to find a password manager you like and that supports passkeys. Try it out on a few less important websites until you feel comfortable with them. If you use multiple platforms, a cross-platform password manager is the way to go.

    FWIW, I had similar concerns in the beginning, but now I really like them.

    💬 1🔄 0⭐ 0
  • Oct 14, 2026, 6:32 PM

    @gavin57

    Finding a password manager you like that supports passkeys won't do you any good because the server requires remote attestation and your password manager isn't on the server's approved list.

    If that hasn't been an issue for you yet, consider yourself temporarily fortunate. Remote attestation is part of the spec for a reason, and that reason is to ensure you are using a “secure” proprietary device.

    @rzeta0 @ScottHelme

    💬 1🔄 3⭐ 1
  • 💬 0🔄 0⭐ 1
  • 💬 0🔄 1⭐ 1
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 14, 2026, 7:36 PM

    @poetaster @rzeta0 @ScottHelme Well, as Passkeys only work on the domain they were created on, it’s difficult for someone to hand their credentials to a fraudulent website because the passkey won’t work.

    I’m not sure how you would socially engineer someone to give up their private key. You would need to compromise the device or the whole password manager account.

    💬 0🔄 0⭐ 0
  • Oct 15, 2026, 5:56 AM

    @gavin57 @rzeta0 @ScottHelme I lived in Calgary when Mitnick walked out of the Nortel offices with Dat tapes handed to him directly. It is quite common for naive users to be led through a convincing corpo web flow to do precisely that private key compromise. I only use hardware devices that are effectively scanners. Some banks (Deutsche, Ing) still offer them. Not impervious to abuse but limited in scope and independant from other services.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 6:01 PM

    @rzeta0 @ScottHelme Basically the same reason why we don't use password based SSH authentication but keys. But for the web it's all "oh no so terrible" for whatever bullshit people make up. WebAuthn is the best thing that happened to the Web in decades.

    💬 1🔄 0⭐ 0
  • Oct 14, 2026, 6:34 PM

    @ljrk

    SSH keys don't have remote attestation and therefore don't force you to use a proprietary device.

    💬 1🔄 0⭐ 0
  • Oct 14, 2026, 7:49 PM

    @argv_minus_one I use open source passkey managers exclusively. The remote attestation part is purely for orgs that want to enforce that their employees use an $org_blessed keychain. As a regular user that isn't relevant for me at all.

    With SSH keys you could (and arguably should) do something slightly related: Have SSH keys part of your CA (SSH certificates) and hand them out through the TPM/SE or a Security Key.

    So a) that remote attestation part is irrelevant to the case aa you (or your org) and choose to not use it and b) the actual major difference/upside of Passkeys vs. Passwords is the same as with SSH keys.

    💬 1🔄 0⭐ 0
  • Oct 15, 2026, 10:09 AM

    @ljrk @argv_minus_one you’re assuming big tech won’t try to force attestation on their services, which historically is the exact opposite of what they do. You already can’t use a lot of android apps unless the hardware and software are blessed by google, and that’s a supposedly “open” mobile OS. People were objecting to that too, but they implemented it anyway. Why wouldn’t passkeys be the same? Unless this shit is out of the standard, it will be used. Just because we’re being eased into it doesn’t mean anything.

    Also RE: using certificates as SSH keys - it’s not similar at all and it’s irrelevant to the discussion.

    💬 1🔄 0⭐ 0
  • Oct 15, 2026, 10:35 AM

    @nrab @argv_minus_one Okay, so in what way is not using Passkeys helping your case? Except for making you less secure?

    Also, by your own words, Google "SafetyNet" etc. is already everywhere. And it makes much more sense for vendors/big corp to attest application state than auth state. They don't gain anything through the latter, while they (from their perspective) arguably gain something through that kind of control over apps. Not all remote attestation is equal.

    It reaches a similar goal.

    💬 1🔄 0⭐ 0
  • Oct 15, 2026, 10:37 AM

    @ljrk @argv_minus_one it helps in the sense that as long as people don’t start falling for passkeys, it’s one less door open to regular people not owning their own devices. Just because big tech already is deploying surveillance and tries to incorporate lock-in, doesn’t mean we should welcome standards with literally vendor lock-in as 1st class citizen with open hands.

    💬 1🔄 0⭐ 0
  • Oct 15, 2026, 10:43 AM

    @nrab @argv_minus_one They can already do RA on a different level (and do so). There's no incentive to do RA for auth for big tech, but if you wanna fight that fight for you, sure. I'd rather have my friends not get phished.

    💬 1🔄 0⭐ 0
  • Oct 15, 2026, 3:46 PM

    @ljrk

    False.

    Currently, there is absolutely nothing stopping me from logging into my bank using the device of my choice. They need a password, and they need to be able to send me SMS messages, and that's it. No RA anywhere.

    If passkeys become ubiquitous, I will be locked out of my bank account and everything else until I get an approved proprietary device.

    If you're more worried about phishing than you are about this, then your priorities are severely misplaced.

    @nrab

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:05 PM

    @ScottHelme I got my phone exposed because of a disguised prompt to authenticate Google. The attackers fooled me into setting up a passkey on my device and they were able to access my Gmail, sending dozens of emails to further infect other potential victims' mobile devices.

    They are relying on pure muscle memory, hoping someone will authenticate from sheer force of habit.

    So I am back to password + 2 factor.

    💬 0🔄 1⭐ 5
  • Oct 14, 2026, 3:50 PM

    @ScottHelme You start

    > The author calls this a "hidden dial" and argues that sites will slowly turn it up,

    and then go on to say.

    > If a consumer website decided tomorrow that it would only accept attested passkeys, it would lock out almost all of its own users, instantly. What website is going to do that?

    This is the opposite of what the original post proposes - instead of a "hidden dial" you're characterising this as a switch to be flipped, and then arguing against that characterisation.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:55 PM

    @ScottHelme Something like the following sequence of events seems plausible, given how we know these actors have behaved in the past:

    1. Google rolls out support for creating attested passkeys in Chrome, while swearing blind it will never require them.
    2. Time passes. Google already has a programme for people with high-value accounts. It offers them a new setting, "Only trust passkeys attested as generated by Google".

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:55 PM

    @ScottHelme

    3. Time passes. Hailing the success of this programme, Google rolls it out as an option for all accounts.
    4. Time passes. Google sends press release that claims to see a large drop in account takeovers, announces that all new Google accounts will require attested passkeys, existing accounts are unaffected.
    5. Time passes. Google sends second press release, announcing passwords for accounts are deprecated, everyone has 6 months to move to attested passkeys.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:56 PM

    @ScottHelme

    6. Time passes. Chrome starts showing warnings on third party sites that use passkeys without attestation, starts ranking them lower in search, etc, ...

    Google has form in this area - for example, phasing out the use of passwords in favour of OAuth, starting to require photo or video to confirm captchas, developer verification on Android, etc. But it doesn't have to be Google, Apple fanboys are just as likely to adopt whatever Apple tells them too.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:57 PM

    @ScottHelme What prevents this? Google employees saying they'll never implement it? That's nice, and of zero value given how quick Google is to fire people when they show an ounce of moral fibre. Mozilla's recent form is to resolutely say they won't bend the knee right up until they do. Other browsers with < 1% market share won't have any say, and will see users migrate if they can't login to their apps.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:57 PM

    @ScottHelme You may still be right, and the original "It's a trap" blog post is overblown. But I don't think your rebuttal addresses the core point.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:59 PM

    @nikclayton None of what you just described is impacted by us using passkeys as they are right now, so the whole thing is irrelevant really.

    If Google are going to do that in the future, then they'll just do that in the future 🤷‍♂️

    💬 1🔄 0⭐ 0
  • 💬 0🔄 1⭐ 0
  • Oct 14, 2026, 4:13 PM

    @ScottHelme But... That's the key point made in the original blog post. It's not (broadly) arguing that passkeys are bad given their typical usage right now, it's arguing that the potential for attestation-creep in the future is the problem.

    That's the point that any rebuttal of the original blog post needs to actually rebut, and you appear to just have said that yes, Google, or any other large player in the space can do that.

    💬 1🔄 2⭐ 0
  • Oct 14, 2026, 4:13 PM

    @ScottHelme Is it fair to characterise this as "You and the original author agree that this (attestation-creep) is a real risk, but disagree about the level of risk (relative to the other benefits of passkeys)"?

    💬 1🔄 0⭐ 0
  • Oct 14, 2026, 7:53 PM

    @nikclayton

    I don't think attestation creep is a real risk.

    What would be the objective of forcing everyone to use passkeys with attestation?

    I can't see a logical scenario where is benefit to doing that, so I can't see why anyone would (Google or otherwise).

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 9:51 PM

    @ScottHelme For Google, anything that makes it more difficult to block ads.

    E.g., a passkey that attests you're on a stock Android variant, and/or have specific Google code on the device (e.g., I believe attestation on Android requires using the non-free Google Play Services library).

    Google's happy to wait years. Chrome's Manifest V3 format, crippling ad blockers, appeared in 2021, V2 support was only removed five years later.

    💬 1🔄 2⭐ 1
  • 💬 0🔄 1⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 14, 2026, 7:30 PM

    @adingbatponder @ScottHelme From everything I’m reading here, Scott already understands passkeys. Where does this video fall on the scale of “I’m staking my reputation on it” to “I have a gut feeling that this’ll prove you’re wrong”?

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 6:40 PM

    @ScottHelme

    > If a consumer website decided tomorrow that it would only accept attested passkeys, it would lock out almost all of its own users, instantly. What website is going to do that?

    “Hi! We noticed you're using an outdated, insecure passkey. To protect your security, upgrade to one of the following secure options…”

    …and they all require a certified unmodified proprietary device.

    Please, *please* think through how things can be used for vendor lock-in. Because if they can, they will.

    💬 0🔄 2⭐ 6
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 11:48 PM

    @ScottHelme oh! I remember seeing this earlier. Trying to pitch people, in 2026, on the premise that the majority of US tech companies won't make some dramatic turn for evil, and that the only way to be safe is to make the browser responsible for MORE sensitive functionality, posted in a form that renders like absolute incontinent ass on Dillo, is certainly a take. I hope people take it exactly as seriously as it deserves.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0