Microsoft Outlook Flaws Let Hackers Breach Systems Without Clicking Links

Israeli security researchers discovered two critical memory safety vulnerabilities in Microsoft Outlook, allowing attackers to compromise systems via crafted emails without user interaction.

Microsoft Outlook Flaws Let Hackers Breach Systems Without Clicking Links
Photo: Unsplash / @simonbhray (Simon Bhray)

Security researchers Shmuel Ozen and Michael Gorelik from the Israeli cybersecurity firm Morphisec have uncovered two critical memory safety vulnerabilities in Microsoft Outlook that allow attackers to compromise systems without requiring any user interaction, such as clicking a link or opening an attachment.

High-Severity Flaws in Email Processing

The more severe of the two discovered vulnerabilities received a 9.8 out of 10 CVSS severity rating from Microsoft, placing it just below the maximum possible threat level. During a controlled demonstration, the researchers transmitted a specially crafted email message. Merely viewing the message within the Outlook preview pane without opening it in a separate window was sufficient to trigger a localized disruption, causing an on-screen window displaying the word "Hacked" to appear.

While the displayed pop-up window was harmless, it proved that the researchers successfully subverted normal software execution and compelled Outlook to perform unauthorized operations using only an incoming email. Such flaws could theoretically enable threat actors to execute arbitrary malicious code, extract sensitive credentials, deploy spyware, or install ransomware.

"He can see your emails, take your passwords on that computer, read the documents you have and transfer them to a second party," Ozen explained in an interview. Within corporate environments, such an infiltration could expose sensitive business, financial, or security intelligence.

Memory Management Errors

The research originated from previous investigations into whether attackers could compromise systems after already obtaining user login credentials. The team shifted to a more challenging scenario: attempting to penetrate a system without possessing any victim authentication details. By generating various email structures and monitoring how Outlook processed them, the researchers noted that the application crashed during one experiment. A software crash often serves as the initial indicator of deeper programming flaws.

Surprisingly, the vulnerabilities resided within the S/MIME digital signature verification mechanism designed to authenticate message origins and protect data integrity. To validate digital signatures, the application must decode attached data. The researchers identified two distinct memory management errors during this decryption phase:

  • Use-After-Free Flaw: Outlook releases a memory block but retains a pointer to its old address, permitting subsequent operations to corrupt newly assigned data in that same memory segment, alongside a secondary attempt to release the identical memory region known as a double free.
  • Uninitialized Memory Assignment: The software creates a new object without properly defining its target memory address, causing it to default to a stale memory location and overwrite critical data in incorrect sectors.

Broader Implications and AI Acceleration

Both defects belong to the broad category of memory safety vulnerabilities. Microsoft classified the flaws as capable of enabling remote code execution and noted that the highest-severity vulnerability requires zero user interaction. The attack methodology mirrors advanced espionage frameworks like Pegasus by exploiting vulnerabilities in legitimate software already running on a device rather than tricking users into installing malicious payloads. Microsoft published security patches for both vulnerabilities in September and October 2026, and no evidence indicates that these flaws were exploited in real-world attacks prior to disclosure.

Ozen noted that artificial intelligence tools significantly accelerate the timeline required for researchers to comprehend communication protocols and formulate attack scenarios. However, these same capabilities are accessible to malicious actors, intensifying the ongoing race between defensive researchers and threat actors. Although standard digital hygiene practices remain essential, the Morphisec research demonstrates their boundaries, proving that systems can remain vulnerable even when users execute all precautions correctly.

Related News