<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xml:base="http://blog.teledyn.com" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
 <title>TeledyN - Blog-Comment SpamBots - Comments</title>
 <link>http://blog.teledyn.com/node/1358</link>
 <description>Comments for &quot;Blog-Comment SpamBots&quot;</description>
 <language>en</language>
<item>
 <title>A mod_rewrite defense</title>
 <link>http://blog.teledyn.com/node/1358#comment-2758</link>
 <description>&lt;p&gt;A &lt;tt&gt;mod_rewrite&lt;/tt&gt; defense against comment-spam  found on &lt;a title=&quot;October 24, 2004 - 11:56&quot; href=&quot;http://drupal.org/node/6663#comment-18700&quot;&gt;Drupal.org&lt;/a&gt;, posted by candygenius ...&lt;/p&gt;
&lt;blockquote&gt;&lt;p&gt;&lt;cite&gt;They are using indexers to find the comment links. I blocked the indexers as well as the originating spammer and cut down on the spam 99%. This block in .htaccess got rid of the latest and most persistent one I have seen.&lt;/cite&gt;&lt;/p&gt;
&lt;pre&gt;
RewriteCond %{HTTP_REFERER} 12\.163\.72\.13 [NC,OR]
RewriteCond %{HTTP_USER_AGENT} (Fetch\ API\ Request) [NC,OR]
RewriteCond %{HTTP_USER_AGENT} (Microsoft\ Scheduled\ Cache\ Content\ Download\ Service) [NC,OR]
RewriteRule .* - [F]
&lt;/pre&gt;&lt;p&gt;&lt;cite&gt;Doesn&#039;t stop Google or anyone legitimate from indexing.&lt;/cite&gt;
&lt;/p&gt;&lt;/blockquote&gt;
&lt;p&gt;That code didn&#039;t work for me &lt;i&gt;directly&lt;/i&gt;, but once I removed the &lt;tt&gt;NC,&lt;/tt&gt; in each line, it worked just fine.&lt;/p&gt;
&lt;p&gt;The same basic technique is applicable to a whole range of abuse-thwarting Apache rewrite rules; I picked up a list at &lt;a href=&quot;http://diveintomark.org&quot;&gt;Dive Into Mark&lt;/a&gt; to detect and defeat all sorts of suspect referrers and User-Agents with no business being in your site (spambots/spybots/offline downloaders) -- the list was probably overkill, banning long since forgotten offenders (I expect they evolve like flu virii) but still instructive enough to use as a template for blocking today&#039;s versions.&lt;/p&gt;
</description>
 <pubDate>Wed, 08 Dec 2004 16:26:44 -0500</pubDate>
 <dc:creator>mrG</dc:creator>
 <guid isPermaLink="false">comment 2758 at http://blog.teledyn.com</guid>
</item>
<item>
 <title>Blog-Comment SpamBots</title>
 <link>http://blog.teledyn.com/node/1358</link>
 <description>&lt;p&gt;Ok, it&#039;s real and it&#039;s official and it&#039;s here today: Automated blog-comment spambots are a reality. &lt;/p&gt;

&lt;p&gt;Today I was hit by two test posts, and the odd grammar of it made it simple to &lt;a title=&quot;Feedster Search: \&quot;here is a free tip\&quot;&quot; href=&quot;http://www.feedster.com/search.php?hl=en&amp;amp;ie=UTF-8&amp;amp;q=%22here+is+a+free+tip%22&amp;amp;btnG=Search&amp;amp;sort=date&quot;&gt;test in feedster&lt;/a&gt; where I found the same odd and pointless test &lt;em&gt;many times&lt;/em&gt; ... right down to the exact same mispelled &lt;span class=&quot;caps&quot;&gt;URL &lt;/span&gt;used as the spammer&#039;s homepage link.&lt;/p&gt;

&lt;p&gt;I posted the following to another victim&#039;s comments, but it bears repeating:&lt;/p&gt;

&lt;blockquote&gt;&lt;p&gt;&lt;i&gt;Looks like we have cause to alert people -- I was hit twice and Feedster shows a few others were hit too, and since the message is almost identical and I&#039;ll bet the &lt;span class=&quot;caps&quot;&gt;URL &lt;/span&gt;too (in the one&#039;s I got, the &lt;span class=&quot;caps&quot;&gt;URL &lt;/span&gt;didn&#039;t resolve, so Verisign picks it up)&lt;br/&gt;&lt;br /&gt;
And no, I don&#039;t see the value except for one thing: As a marker.  I have heard that spammers who find open comment blogs will leave a marker that can be found with a subsequent search (Feester?) and that may explain why the previous blog-comment spam I&#039;ve received always goes to the same three blog posts.  I used to edit their comments to remove the &lt;span class=&quot;caps&quot;&gt;URL&lt;/span&gt;s and then ridicule them, but maybe this is a mistake, maybe they just want that highly recognizable string to stay there so they can find it later ... or (as I believe is the case today) where robots can find them.&lt;br/&gt;&lt;br /&gt;
I got another today, simply some guy&#039;s name &quot;rules&quot; with a link to the same name dot com, nothing more, and that link was bogus as was the name@aol ... it &lt;em&gt;must&lt;/em&gt; be some sort of marker they are leaving to bootstrap some other planned deployment.  That&#039;s the only explanation I can think of that makes sense.&lt;/i&gt;&lt;/p&gt;&lt;/blockquote&gt;

&lt;p&gt;end of an era here folks.  Open blog comments: 1997-2003. &lt;/p&gt;

&lt;h4&gt;A Proposal for Blogcrafters&lt;/h4&gt;

&lt;p&gt;I&#039;ve &lt;a href=&quot;//mt/archives/001235.html&quot;&gt;written previously&lt;/a&gt; on the problems with some of the obvious solutions such as the bubble-obscured pass key; any viable solution has to be open to all platforms, including those used by blind surfers and low-tech browsers.  The solution has to preserve the casual conversation, &lt;i&gt;in situo&lt;/i&gt; to retain the forum sense of a thread, and it has to thwart the blog spammers by making each comment somehow &lt;em&gt;accountable&lt;/em&gt; ...&lt;/p&gt;

&lt;p&gt;I propose an &lt;span class=&quot;caps&quot;&gt;HTTP&lt;/span&gt;-based comment moderation or author confirm scheme.&lt;/p&gt;

&lt;p&gt;Just as we do for automated subscription confirmations, I propose we still allow our blog engines to accept posts from anyone, but &lt;em&gt;before the post is displayed&lt;/em&gt; someone must confirm the message.  The blog engine might send the comment back to the author with a hash-code &lt;span class=&quot;caps&quot;&gt;URL &lt;/span&gt;they must click to have the comment confirmed and displayed.  &lt;/p&gt;

&lt;p&gt;Blog owners would still retain the power to override this comment-hold manually through the control panel, or perhaps they too are sent the email and can confirm or reject it. A whitelist of trusted parties wouldn&#039;t be sufficient since the spammer could just borrow emails they find on the same thread of comments.&lt;/p&gt;

&lt;p&gt;Since many, probably most webhosts can&#039;t accept email, email reply to confirm is probably impractical. The simplest hack for a blog like MovableType would be a comment setting for moderated and/or confirmed comments (maybe this already exists?) where email alerts it already sends on new comments are extended to include the confirm (or instant delete?) &lt;span class=&quot;caps&quot;&gt;URL &lt;/span&gt;in the message.&lt;/p&gt;

&lt;p&gt;Is this possible as an mt-plugin?  Looks like another job for &lt;a href=&quot;http://www.lazyweb.org/&quot;&gt;Lazyweb&lt;/a&gt;&lt;/p&gt;</description>
 <comments>http://blog.teledyn.com/node/1358#comments</comments>
 <category domain="http://blog.teledyn.com/taxonomy/term/2">here comes everybody</category>
 <pubDate>Wed, 01 Oct 2003 17:07:29 -0400</pubDate>
 <dc:creator>mrG</dc:creator>
 <guid isPermaLink="false">1358 at http://blog.teledyn.com</guid>
</item>
</channel>
</rss>
