> the from: address in every email is an arbitrary and unverified text string that the sender provides

DKIM et al came back clean.

As for URL scheme, I mean the format and layout of URLs - because it was an MITM attack, they matched 1:1.

how did you evaluate the sender address via DKIM to get "clean" response? I mean I know there are methods to verify stuff about a received email, DKIM by itself only handles message integrity and not sender details, for that you need to fold in DMARC -- but there are all WILDLY technical details that are certainly not what anyone is gonna do before clicking a link in a message body

> As for URL scheme, I mean the format and layout of URLs - because it was an MITM attack, they matched 1:1.

"scheme" is a well-defined domain term that refers to the e.g. `https://` part of a URL/URI -- but that aside, I still don't get what you're saying here? what is "format and layout of URLs" and how does that relate to "mitm attack"?

to cut to the chase, a malicious email maybe contains a link, to a URL, that a victim can click on. but if that link says it goes to `https://npm.org` then it actually does go to `https://npm.org` and there isn't like any special secret way for an email to hijack or mitm that domain or URL resolution. if the link is actually `https://npn.org` then that's a totally different thing, it's not a mitm attack, there is no concept of "format or layout" of that totally different URL "matching 1:1" with `https://npm.org` -- unless you're talking about something totally different to what I'm understanding?

edit: wait are we talking about an email sent from a domain `npmjs.help`? DKIM and DMARC and URL scheme validation don't even enter the picture here, this was no kind of mitm attack by any definition -- "npmjs.help" is clear-as-day a malicious domain, and any email from it a clear-as-day phishing attempt.. ! it's fine, we're all human and etc. but it just underscores the issue here being minimizing blast radius of failures, and not anything related to any specific user/human

▲

junon a day ago | parent [-]

I think you are missing a lot of information I've posted elsewhere in this thread and the original HN post. I didn't minimize anything; I would hope most agree that if anything I've maximized the message as much as I possibly could to prevent further damage.

1. My email client does the validation of certain integrity and security checks and shows a checkmark next to senders that pass. Since npmjs.help was a domain legitimately owned by the attackers, it passed.

2. The link in the email lead to their site at the same domain, most likely performing a MITM between my browser and npm's official servers.

3. You're arguing semantics about "scheme". Please try to understand what I'm attempting to convey: The URLs appeared to match the official npm's site. There was no <a href> trickery. Once I had it in my head (erroneously) that .help was fine, nothing else about the attack stood out as suspicious when it came to the URL or domains.

4. Emails themselves are not MITM attacks, no. I didn't respond to an email with my credentials. I would never do that. But that isn't what I've ever claimed to have happened.

5. The URLs being similar or identical to npm's isn't how they technically achieved the MITM. The URLs being similar was to avoid arousing suspicion.

Hopefully that's explanatory enough.

	▲	kiitos 17 hours ago \| parent [-]
		> The URLs appeared to match the official npm's site. The domain "npmjs.help" is pretty clearly malicious at a glance, just from the ".help" TLD alone, but yeah as you say > Once I had it in my head (erroneously) that .help was fine, nothing else about the attack stood out as suspicious when it came to the URL or domains. well except that presumably you clicked on a npmjs.help link and the new tab ended up at npmjs.com? but yeah it's a tough break, don't mean to needle you, hopefully learning experience