Something very surprising - while searching for Windows Live Writer Installer on Google Search, one of the top search results directly links to Writer.msi - an executable Windows binary file stored on one of the Microsoft webservers.
Now since this file is hosted on Microsoft, it maybe pretty safe do download and run it on your hard disk but could there be a situation when Google indexes executables that are trojans or even viruses ?
You click on the search result and if the browser download manager is configured to download files automatically, you could have a potentially unsafe file sitting on your file system that may spell trouble the moment someone double clicks that file.
The filetype search operator doesn't work for msi or exe files extensions but inurl works. For example, we found a couple of exe files on Sourceforge.net appearing in Google search results.
Not sure how Googlebot reacts when it encounters binary files like exe's or msi's but this issue might be something to worry about.
Find this article at: http://labnol.blogspot.com/2007/01/this-is-scary-googlebot-indexing.html
web: http://www.labnol.org/ email: amit@labnol.org
Reader Comments
Scary eh?!
Written on 10/1/07 3:27 AM
The filetype operator works pretty well for MSI files too. When I tried to search the term 'download' with filetype MSI it returned quite a lot of results. http://www.google.com/search?hl=en&q=download+filetype%3Amsi+
But I don't see a potential risk here unless the user is novice and doesn't have up-to-date AV definition.
Written on 10/1/07 3:40 AM
I noticed the same case with EXE file too.. Google is indexing EXE's too..
Written on 10/1/07 3:41 AM
:), do not get surprised. Google has a good doc for it. http://www.google.com/help/faq_filetypes.html.
Look at point 9, it will be helpful.
Written on 10/1/07 3:07 PM
Amit,
Google has tied up with Websense and they filter out all Trojans and other sites of the search results. Also if a trojan is found on some particular site and if you link to it, your ranking will be pushed down.
I am not sure how effectively it has been implemented but positive that they are designing such a system.
Written on 11/1/07 10:12 AM