Re: Re: [WEB SECURITY] Universal XSS with PDF files: highly dangerous
YMMV, but in our own testing of server-side defenses for Apache, we had to
resort to the following to extract the anchor information:
# mod_rewrite defense
RewriteCond %{THE_REQUEST} .*\.pdf[^A-Za-z0-9._?&%-]
RewriteRule (.*\.(pdf)) $1 [R,L]
# mod_security defense
SecRule REQUEST_URI_RAW: ".*\.pdf[^A-Za-z0-9._?&%-]" \
"deny,log,status:502,id:951004,severity:2,msg:'DOM attack on static
content'"
In the case of mod_rewrite, I had to use THE_REQUEST, which captures the entire
1st line of an HTTP request, not just the URI (hence capturing the anchor); in
the case of mod_security, I think I just got lucky, I can't explain why
REQUEST_URI_RAW worked while REQUEST_URI did not. This was tested with Apache
2.0.59 and modsecurity-apache_2.0.4.