TeamViewer Flaw Let Helpdesk Hosts Attack the Technician
By @sharedot · · 8 pages
- Cybersecurity
- Teamviewer
- Cve 2026 111585
- Vulnerability
- Remote Access
A high-severity TeamViewer flaw (CVE-2026-111585) let a malicious host crash the technician's viewer or run code, fixed in version 15.83.
What Happened: The Support Session Turned Weapon
TeamViewer has patched CVE-2026-111585, a high-severity flaw (CVSS 8.8) that reversed the usual risk model of remote support: the machine being controlled could attack the computer that connected to control it.
In a TeamViewer session the connecting side is the viewer — typically an IT support technician — and the controlled side is the host.
Once a technician accepted a session to an attacker-controlled or compromised host, the attack required no further clicks and no password theft, working in TeamViewer's default configuration.
The flaw affects the Full Client on Windows, macOS and Linux, the Windows Portable edition, and the Android Remote Control app in all versions before 15.83, including September's 15.82 security update.
Why It Is Surprising: The Attacker Is the Host
As Remedio researcher Omri Dar wrote, "the attacker is the host and the victim is the viewer."
A malicious host triggers the bug from inside an ordinary, established session using only messages the protocol already supports, and can force the vulnerable code path simply by refusing to negotiate AV1 and H.264 video codecs, pushing the viewer onto its bitmap decompression route.
CyberSecurityNews notes this places the technician's own workstation — often holding credentials and tooling — directly in danger, and that session approval alone is no protection.
The Evidence: A 32-Bit Overflow in Screen Decode
The flaw is an integer overflow in how the viewer sizes its drawing surface. The host announces screen dimensions via a DisplayParams message, and the viewer allocates a CBitmapBuffer using 32-bit multiplication.
Remedio showed that dimensions such as 32,768 by 32,768 pixels at 32 bits per pixel overflow the calculation, wrapping to roughly 4 MB of memory instead of the more than 4.29 billion bytes required.
The decoder validates incoming rectangles against the declared dimensions rather than the actual allocation, so pixel data spills past the buffer in a heap overflow.
Remedio demonstrated code execution on Windows, bypassing Data Execution Prevention and Control Flow Guard to redirect a call to WinExec — though its demo used a host-side hook for part of the attack data, so it was not yet an attack carried entirely over the network.
The Stakes for Helpdesks and IT Teams
TeamViewer is widely used by corporate help desks to troubleshoot employee devices, so technicians routinely connect to machines they do not trust. At minimum the bug gives a malicious host a repeatable denial-of-service, crashing the viewer's application on demand.
On Windows, successful exploitation could run arbitrary code with the viewer process's permissions — CyberSecurityNews notes that this does not automatically mean SYSTEM or root access, but a technician's user session is still a rich target.
TeamViewer's bulletin TV-2026-1011 also covers two other high-severity flaws fixed in 15.83: CVE-2026-92372, an out-of-bounds write in screen-image and .tvs recording decoding on Linux and macOS (CVSS 8.8), and CVE-2026-92374, a TOCTOU symlink flaw letting a local attacker gain root on macOS (CVSS 7.8).
A companion bulletin, TV-2026-1012, patches six local privilege-escalation flaws in Remote Management add-ons.
Patch Now, Don't Rely on Session Approval
TeamViewer shipped the fix on October 27 in version 15.83, which covers the Full Client, Host, Portable and Android Remote Control App, with fixed maintenance builds available for legacy versions.
TeamViewer lists no workaround, so updating is the only mitigation — and The Hacker News notes that last month's 15.82 update did not remove this flaw. Administrators should also update the Remote Management add-ons to their latest versions.
TeamViewer said it was not aware of any exploitation in the wild when the fix published, the flaw is not in CISA's Known Exploited Vulnerabilities catalog as of October 29, and Remedio released no exploit code.
Disclosure Timeline and What Comes Next
Remedio notified TeamViewer's Product Security team on August 14, 2026, and TeamViewer confirmed the issue on October 22, assigning the CVE-2026-111585 identifier with a CVSS score of 8.8. The patch arrived October 27, and Remedio published its technical write-up on October 29.
The open question now is whether the demonstrated Windows exploit path — which per CyberSecurityNews used a host-side hook to place corrupting bytes rather than a fully network-carried payload — can be completed remotely, and whether similar code execution is achievable on macOS, Linux or Android, which TeamViewer says depends on the platform.
For support teams, the immediate priority flagged by CyberSecurityNews is updating viewers before connecting to untrusted computers.
Sources
- securityweek.com › TeamViewer Flaw Could Let Malicious Hosts Take Over Support Technicians' Machines
- thehackernews.com › TeamViewer Flaw Lets Remote Computers Attack Users Who Connect to Control Them
- hackread.com › TeamViewer Vulnerability Lets Attackers Run Code on IT Support Staff's PCs
- cybersecuritynews.com › TeamViewer Security Flaw Could Turn Remote Support Sessions Into Computer Takeovers