Google Webmaster Tools is reporting crawl errors for links that don't exist

Question

In Google webmaster tools > diagnostics > crawl errors I'm seeing a bunch of errors that link to sub-pages like /tour/overview/Scripts/ForceDownload.asp The /tour/overview does exist, but this subfolder doesn't (and shouldn't - it's not even an ASP server). More bafflingly, the "linked from" pages listed do exist but do not have these links.

I suspect some malware may be the cause but I've yet to find any trace of it.

Answer 1

Looking at the source of your page, I see the following lines of JavaScript code:

//call multiBox
var initMultiBox = new multiBox({
    ...
    addDownload: false,//do you want the files to be downloadable?
    pathToDownloadScript: './Scripts/ForceDownload.asp',//if above is true, specify path to download script (classicASP and ASP.NET versions included)

Apparently, Google is picking up this path from the JS code and trying to crawl it.

The ASP script and the JS code appear to be part of multiBox. By the way, you're probably lucky that your server doesn't run ASP — looking at the source, that script seems to allow downloading arbitrary files under the webserver document root, bypassing any server access controls. (The ASPX version is slightly better, in that it comes with a forbidden extension list, but it still seems to treat any unrecognized extensions as allowed.)