A hosted IMAP workflow that gives mobile and desktop mail clients one explicit spam-confirmation action, then applies the resulting block decisions from centralized state.
Scope note: this page documents the implemented workflow and safety boundary without publishing mailbox data, sender data, credentials, infrastructure details, or private operational records.
A normal Spam action on mobile and desktop clients did not reliably create the same server-side Junk signal. That made it unsuitable as a single, cross-device instruction for a background worker.
The goal was not to replace mail filtering. It was to provide one deliberate operator action that works from either device type and is processed even when personal computers are off.
The workflow uses a dedicated IMAP folder named Spamwatch Confirmed. Moving a message there is an
explicit confirmation, independent of how a client represents its normal Spam or Junk action.
On its five-minute schedule, the hosted worker processes confirmations through centralized MySQL state. Confirmed spam is permanently deleted by design, and later mail from a confirmed blocked sender is removed from the relevant mailbox folders.
Owned email domains are protected from being learned into the blocklist. A message that merely presents an owned domain therefore cannot poison the sender or domain rules, even when its confirmation message is deliberately removed.
This is a mailbox automation workflow, not a claim that phishing is prevented. It does not verify sender identity and cannot block all phishing; it acts only on explicit confirmations and stored block decisions.
The current configuration deliberately sends no spam-notification emails. Instead, logs and cron output provide operational evidence of each run and its aggregate confirmation and deletion activity.
Centralizing live state avoids conflicting decisions between devices while keeping the worker independent of a particular computer being online.