The Web sites of three of the security industry's best-known companies include security flaws that could be used to launch scams against customers, according to a new report.

The report, from security watchdog site XSSed, verified 30 cross-site scripting (XSS) vulnerabilities across the sites of McAfee, Symantec and VeriSign. The flaws could be used to launch scams or implant malicious code on the systems of visiting users, according to XSSed.

Recent research has shown that attackers are increasingly - even predominantly - now using legitimate sites to host their malware, a tactic that makes the malware distribution sites more difficult to shut down.

XSSed's results show that even major security firms are not exempt from the problem, according to XSSed.

In January XSSed found that 60 Web sites that had received a "Hacker Safe" certification from McAfee's ScanAlert service were in fact vulnerable to XSS attacks.

McAfee and other major security firms have downplayed the seriousness of XSS flaws, compared for instance to flaws that allow an attacker direct access to customer data stored on a server.

In recent months the real-world exploitation of XSS flaws has boomed, exploiting major Web sites such as MySpace, Paypal and a major Italian bank.

Last week ScanSafe reported that 68 percent of all malware it blocked in May was found on legitimate sites that had been hacked, more than quadruple the level of a year earlier.

Such flaws can be used to steal user cookies, to steal website login credentials and to exploit users' trust of a site in other ways, and in theory can be shut down quickly once the owner of the site is made aware of the problem.

However, the techniques used by hackers are highly automated, allowing them to "colonize" large numbers of vulnerable sites at once, ScanSafe noted. By contrast, the fixes are not necessarily so easy, researchers have noted.

In a research note in May, F-Secure noted that one legitimate site had been repeatedly hacked and used to spread malicious code, and each time it needed to be contacted to fix the problem.

"The site cannot simply be pulled offline without collateral damage to the legitimate business. So the website's administrator must be contacted to repair the damage," said F-Secure's Sean Rowe in the research note.