SPAM control for podmins; could someone summarize please?

I have tried to search for a place which actually summarize what antispam possibilities a podmin have.

So, I have seen several threads, most of them from many years ago, mentioning various spam problems:

  • how to control spammy registrations on the pod
  • how to control fake email for the spammy accounts (how not to send email with unverified addresses)
  • how to control connections and content of spammy pods
  • how a podmin can destroy a spammy acocunts’ contributions (posts, comments, likes, whatever) apart from closing the spammy account
  • is there any way to control actual spam by registered accounts (eg. rate limiting, watched keywords, etc.)

(This question was partially arisen by spammy local registrations, and partly due to public post relays.)

1 Like

No? No. I thought so.

Anyway. Here are my assorted ideas and tools.

Admin interface

  1. There should be a time range by registrations list of users, just like “weekly users stats”, since that’s the way I could review recent registrations (spambots).

  2. This list, as well as “Weekly user stats” should be in table form. The table would contain the most relevant data helping the admin make the distinction between spambot and user:

    • username - register time - sign in times - last seen - email - last ip
      Hopefully sortable on the fields.
  3. All rows shall contain small buttons for close, lock or wipe (this one would delete all the posts of the spammer), and a checkbox. Also a link to view the account and its posts (no javascript so it can be opened in background tabs).
    By the way following and followers would be useful as well.
    At the bottom there would be close all checked etc. buttons.

  4. Admin shall have a “delete current userbutton in the header to quickly able to lock/close/wipe the user s/he is observing. Or a button on the side. Wherever.

Automated means

Right now I use two scripts to clean up the trash.

  1. One gets the subnets of known spambots and removes all account registered from those subnets.

  2. The main script collects accounts where there was only one login and it’s older than a few days or weeks. I close and wipe these.

  3. …also collects accounts where there were a few logins only within a short timeframe and it’s been a longer time ago (for me it’s 3 months). I usually close these.

Both scripts work from the postgres directly.

I would be glad if these could be automatised, maybe if I had a framework which only needed to have a few specific lines of code inserted. A working example for these…

2 Likes

You see the Anwesens - diaspora lacks a full set of spam controls:

  • on registration emails are not checked
  • Pods can not be blocked
  • users can not be closed

But there is hope:
Community woke up (“woke”… uh…) and has made some contributions:

Still not finally reviewed but a valid pull request to to core code:

block a pod

Enhance reports

#Block / Close an account

You might easily merge this branches to your code - but without any warranty.
Share your experience!

1 Like