The unusual bug, in Microsoft anti-malware software such as Windows Defender, could be exploited without the recipient even opening the message.

Researchers working for Google's Project Zero cyber-security outfit discovered the flaw at the weekend.

The fix has been specially pushed out hours before the software giant's monthly Tuesday security update.

Hackers could exploit the flaw simply by sending an infected email, instant message or getting the user to click on a web browser link.

Windows 8, 8.1, 10 and Windows Server operating systems are affected by the bug.

Anti-virus software such as Windows Defender would merely have to scan the malicious content for the exploit to be triggered.

On some computers, scans are set up to occur almost instantly - "real-time protection" - or to take place at a scheduled time.

"Anti-virus normally tries to intercept these things before you get to them," said cyber-security expert Graham Cluley.

He added it was "tremendous" that Microsoft had released the patch so quickly.

The bug was discovered by Google Project Zero researchers Tavis Ormandy and Natalie Silvanovich.

And Mr Ormandy later tweeted he had been "blown away" at the speedy response.

The vulnerability allows for remote code execution: "the thing all the malicious attackers are aiming for", Mr Cluley told the BBC.

"It means they can install code on to your computer without your permission - it means they can hijack your computer."

Mr Cluley did add, however, that he thought the Project Zero protocol for announcing the vulnerability had been risky, because it included information that malicious hackers might have found useful.

"That can help the bad guys," he said.

Windows users can check that they are running the latest Windows Defender engine version (1.1.13704.0), which should download automatically, to make sure they are not at risk - or hit the update button.