Security researcher Jeremy Conway says he has discovered a way to spread malicious code across PDF documents on a victim's computer. The attack leverages a flaw in the way the PDF file format works, adding malicious data to legitimate PDF files that could then be used to attack anyone who opens them.

Conway, a product manager with NitroSecurity, had already developed a technique for injecting the malicious commands into PDFs. But his attack only worked when there was some other malicious program on the system that added the code. That all changed last week, when researcher Didier Stevens showed how a PDF document could be altered to run an executable file on a victim's computer.

"When I saw Didier's hack, it was the first time I could do it from completely inside the PDF," Conway said.

Hackers have known for some time that PDF readers could be manipulated in this way, but Stevens' attack showed how one reader -- Foxit Reader -- could launch the executable without even notifying the user. Foxit has now patched this bug, but the underlying flaw in the PDF standard can't be fixed without changing the PDF standard itself.

The link for this article located at Computer World is no longer available.