What's the scope of the vulnerability?
This is a vulnerability that could allow an attacker to run script of his choice on the user's system, via an HTML e-mail. Such a script could take action on the system as though it was the user.
An attacker would only be able to exploit this vulnerability if the recipient has Word configured as the e-mail editor and the recipient chooses to reply to or forward the e-mail.
The attacker's actions would be limited by any restrictions which govern the user's actions. Thus, in an environment where accounts adhere to the rule of least privilege, the attacker might be significantly limited in the actions his program could take.
What causes the vulnerability?
The vulnerability results because of a flaw in how the WordMail editor handles scripting contained in HTML when an e-mail message is replied to or forwarded. In certain circumstances, the scripting is handled in an unsafe manner and could run without warning the user.
What is WordMail?
WordMail is another way of saying that you have enabled Word as your e-mail editor. WordMail allows you to create new e-mail messages using most of the features found in Word, such as formatting, AutoText and Check Spelling as You Type.
How does WordMail handle HTML e-mail?
WordMail is used for composing, replying to, or forwarding e-mail, including HTML e-mail
What's wrong with how WordMail handles HTML e-mail in replies or forwards?
The problem results because e-mails are handled differently when read than when replied to or forwarded. When you receive and view an HTML e-mail, the message is subject to the Internet Explorer security zone settings, which are honored by Outlook. Scripts are not run because of the Internet Explorer security zones.
When you reply to an e-mail or forward the message, the new message is in the same format as the message you received. A reply or forward has been treated as being the same as creating a message, which means Word is in a less-restrictive creation mode that doesn't block scripts, so the scripts could be run.
How might an attacker exploit this vulnerability?
An attacker could create an HTML e-mail containing script, then send it to another user. If the recipient is using Outlook 2000 or Outlook 2002 and chooses to reply to or forward the attacker's HTML e-mail using WordMail, the script could then be run. The script would have access to the user's local system resources and can execute with the same privileges as the user.
How would the attacker know whether the user was using WordMail?
The attacker has no way to discover remotely if a recipient is using WordMail.
Does the vulnerability provide any way for the attacker to force the user to reply to or forward the mail?
No, the user must choose to reply to or forward the e-mail.
Is it possible to craft an HTML e-mail message like this by accident?
No, it would require very specific, detailed knowledge and such a message would have to be specifically constructed with malicious intent.
What can I do to protect myself against this vulnerability?
The best way to protect yourself is to apply the patch to systems running Outlook 2000 or Outlook 2002.
Are there any other steps I can take to protect myself?
Customers who have enabled a new feature feature added in Office XP SP-1 that lets you read all non-digitally-signed e-mail or non-encrypted e-mail in plain text format are protected against attempts to exploit this vulnerability.
The KnowledgeBase article Q307594 describes how to enable this feature.
However, you should still apply the patch in case you disable the read as plain text option in the future.
What does the patch do?
The patch eliminates the vulnerability by having Word handle all forwards and replies in Design mode, which will not allow scripts to be run.