@julian Yes, there is potential for abuse. In my current forum (Drupal 6) I have implemented a few checks and balances:

1. X spam clicks from different users removes the post
2. (X*2)+1 spam clicks across all of the user's posts locks the account
3. Each user gets a maximum of X*2 spam clicks per hour (so a single user account can't do a fast organized attack against a user to get him maliciously blocked.)
4. A user's supply of valid spam reports gradually replenishes.
5. A user's record of spam hits gradually erases itself.
6. Users that have been registered for longer than Y days can not be locked out.
7. Very new users can mark a post as spam and contribute to it being eliminated, but their clicks don't count toward the total needed to block the user's account.

This has worked extremely well, and although it does occasionally require manual intervention, it's much less than the amount of time needed to manually sift through a reported spam moderation queue.