Hey! Google! You have someone inside your Tech Dept., who repeatedly blocks various parts of the AntiMullah site,when something prominent is posted in that section that displeases the Mullahs.
NONE of the seven or eight sections/blogs are SPAM nor have ever been over the years - nor will ever be.
PLEASE take a good look at them and take action to prevent any of these blogs being so easily and UNWARRANTEDLY automatically blocked - and require PRIOR PERMISSION FROM MANAGEMENT for this to happen.
This usually takes place on a weekend when there is nobody from management available to check and unblock them - and since even a cursory glance shows none of these blogs are SPAM, it results as a harassment - from within your organization, using the SPAMbot as an excuse.
If you look at the history of any of the group of AntiMullah blogs, I would posit that there is nothing that could trigger a spambot and it is physically done internally at your night shift tech dept. - always happens come evening time, where there is nobody to verify and release.
For instance the Oghab (Eagle) Movement of General Rouhani that is actively opposing the Mullah regime and the top article (in Farsi) posted below this missive to Google advises people of the existence of this movement. Which has had it's contact phones and faxes blocked for out-going calls from Iran.
The only possibility/excuse for your very SELECTIVE Spambot I can think of is that I have installed three or four different "subscription feed links" like Google, Yahoo, Plain Text etc., and if there are hundreds of responses triggered to subscribers (and I get an aggregate of about a thousand a day) for the various group of blogs under the AntiMullah banner, then you need to make the blogs immune to being blocked for being successful.
AGAIN, there must be a setting for White Listing the Blogs from the Spambot harassment - except if directly approved by someone in Google Management. Not the other way round that the Blog gets blocked till some one at management level arrives at work, sees it was yet again a false positive and clears the block.
Suspend the Block function but have it send a message to a supervisor who can THEN evaluate whether it was validly flagged or another false positive that seems to coincide with Mullah or local political "censorship" by pressing the button.