Bulk-clean spam from the pending comment queue

There is a difference between a comment queue you want to read and one you just want gone. When a month of link spam has piled up behind a genuine trickle, reviewing each entry is not moderation, it is punishment. This is the decisive version: you set the rule, the agent clears the queue in bulk, and because it only marks or trashes, nothing it does is final.

My pending comment queue is buried in spam and I want to clear it in bulk, not
review it one by one. Still, action nothing until I approve the rule.

First, read the pending queue and tell me what I'm dealing with: how many are
waiting, and a quick sample of the obvious junk so I can see the pattern, the
link-stuffed bodies and gibberish authors. Don't mark anything yet.

I'll give you a rule, something like "mark everything with three or more links
as spam and trash the rest of the junk." Apply it in batches, marking as spam
or sending to trash, and show me each batch as you go. Both spam and trash are
recoverable, so a wrong call is something I can pull back. When you're done,
tell me how many you marked as spam, how many you trashed, and what you left
in the queue for me.

Why this is safe to run

Reading the queue and moderating a comment are two separate abilities with two separate switches, so the opening read genuinely writes nothing: you see the size of the problem and a sample of the pattern before setting a rule. You approve the rule before a single comment is touched, and the agent works in batches you watch rather than clearing everything in one silent pass. Every action is a spam mark or a move to trash, both recoverable, so a rule that catches one good comment is a quick undo, not a loss. And comment reads never hand back an email address or an IP, so the sample stays privacy safe.

Already connected? Ask your agent for the “bulk-clean the comment queue” recipe, give it a rule, and it runs the same read-first sweep on your own backlog.