Cyber criminals are targeting the popular Wordpress blogging platforms with password cracking attacks.
Security expert Brian Krebs reported uncovering the campaign, publishing a list of sample WordPress usernames and passwords used in this attack on Saturday.
Security firm Sophos reported that the crooks are stealing the information using a botnet to launch automated dictionary attacks.
"Word from the anti-DDoS world is that a botnet is responsible, with estimates of ‘up to 90,000,' ‘more than tens of thousands,' and ‘up to 100,000' infected computers orchestrating the felonious login attempts," wrote Sophos researcher Paul Ducklin.
Dictionary attacks aim to hack accounts by guessing the usernames and passwords. The attacks can vary in sophistication, with some being based on hard research using data stored on the victim's Facebook page or Twitter profile, while others can simply be random guesses.
"The idea is simple: automate the password guessing, speed up the attack, and don't spend too long on any individual site. Look for the low-hanging fruit, and harvest it as quickly as you can; if you can't get in within a few hundred or thousand attempts, move on to the next potential victim," wrote Ducklin.
"It's doorknob rattling, but on an industrial and international scale."
Ducklin said that the attack targeting Wordpress is a simple one and so users of the sites being attacked that follow good password security, such as using random passwords with a mix of letters, numbers and characters, should be safe.
Zombie botnet networks are a constant problem in the security community, becoming an increasingly common tool in cyber crooks arsenal.
The networks enslave unwitting internet users using a variety of techniques, including phishing messages loaded with malicious web links and PDF attachments.
Security vendors have warned that despite the simplicity of the attacks thousands of people are still falling for the scams.
Most recently Russian antivirus firm Doctor Web reported hijacking control of a botnet infecting over 100 computers per hour.