Noted, thanks for the feedback! Will consider changing the requirements for the password.

2FA is on the roadmap and we’re planning to release it within the next two weeks.

The official NSA advice for several years has been "do not enforce restrictions on characters in passwords".

It's simple to calculate the actual entropy and check against a list of common weak passwords (large lists are easily obtained and kept up to date)

Makes a lot of sense. I guess the only requirement that would still be helpful is minimum password length?

See 5.1.1.1 & 5.1.1.2 on https://pages.nist.gov/800-63-3/sp800-63b.html. Eight minimum, accept at least up to 64 characters, forbid breached passwords, dictionary words, aaaaaaa style passwords, and usernames, but beyond that:

> Verifiers SHOULD NOT impose other composition rules (e.g., requiring mixtures of different character types or prohibiting consecutively repeated characters) for memorized secrets. Verifiers SHOULD NOT require memorized secrets to be changed arbitrarily (e.g., periodically). However, verifiers SHALL force a change if there is evidence of compromise of the authenticator.

Very helpful, ty!

What do you think of password-less sign-ins, e.g. email-based sign in or OAuth 2?

The email-based “magic link” approach I’m seeing a lot more of lately, and I despise it.

I generally find it annoying too, but I think it's reasonable as an option. (Although if it is added as an option, ideally make sure it's possible for password managers to fill in both the email and password in one shot.)

Why do you despise it?

My password manager is faster and easier (and the email flows tend to make that slower, hidden behind a "login with another method" small grey link), email often has deliverability delays, and it's insecure to boot.

I can second that. I prefer password logins over magic links via email. It adds too much friction. I can understand that it might be easier for people who are not using a password manager.

For most services I prefer a combination of password and authy-based 2FA. For very specific purposes, some kind of hardware-based authentication, for example with yubikeys makes sense.

Magic links are great because they essentially eliminate ATO attacks and allow startups to bypass SAML requirements by delegating to the mail provider. I’m 100% willing to forgo efficiency gains of passwords and managers as both a user and developer.

All you're doing is punting your security to my email provider. It's annoying how many companies assume I have absolute trust in my ISP/Google/Miscrosoft/IT.

> essentially eliminate ATO attacks

They kinda increase the blast radius of an ATO attack on your email account, though.

Replying too late, most likely. But with access to your email account someone could reset your password to what they want it to be and gain access.