On 2025-06-15 15:20, Atro Tossavainen via mailop wrote:
I have asked them to uniquely identify sending accounts with some
custom header / token
>
You're asking for something that is already there.

implicit. I want explicit. and I actually would like to propose an SMTP extension, but it is still work in progress and this is my part-part-part-time endeavour (last time I dealt seriously with communication protocols was in the 1990ies).

Goal: per sender encryption-hardened identity; per message encryption-certified purpose header; automated/scalable feedback system for ESP to discipline their senders.

In a nutshell:

recipient at EHLO stage tells sender whether this standard is mandatory, optional, (or ignored)

recipient also indicates what kind of purpose emails are accepted. everything else is not accepted. not sure yet how to implement it at individual mailbox level.

sender adds three headers: (1) a token that together with a well-known location and the sender's FQDN point to the sender's public key. (2) a certified message purpose. (3) a reporting URL.

ESP can already stop sender in track if the purpose is not listed as accepted by recipient.

on recipient's end, the user sees a line asking whether the content of the message is the declared purpose. Yes or No. If a sender sends a "survey" and calls it "transactional" the answer is No. If the sender sends a newsletter and the recipient is subscribed to that newsletter, no surprise, the answer is Yes. Or the end-user ignores, which is fine too, the UI shall be unobtrusive and not waste end-user's time.

The click on the Yes/No reports triggers the reporting URL. ESP can automatically collect and process a sufficient quantity of feedback to police and discipline their senders if necessary.

Of course this will only succeed if enough mailbox operators get on board.

For me personally, I can afford to be hardline and set up my SMTP server to reject everything except what I whitelist. That is not the case for most users. Most users will end up putting expensive (compute-cycles) anti-abuse resources to achieve about the same result. so the question boils down to whether the "federated world" that is email wants to be more efficient or not.


On 2025-06-16 04:37, Atro Tossavainen via mailop wrote:
> You are right, of course. So now somebody at Mailgun knows a few
> spamtraps to delete :-D

security by obscurity has seldom been a winning proposition. A sufficiently large block will convince even the most negligent of ESPs to behave.

Yuv
_______________________________________________
mailop mailing list
[email protected]
https://list.mailop.org/listinfo/mailop

Reply via email to