GIMP CVE-2026-18307

Exploitation requires the target user to open a malicious TIF/TIFF file in GIMP - either by visiting a web page that triggers automatic file opening, or by manually opening a delivered file. The CVSS vector (PR:N) confirms no prior authentication or privileges are needed on the target system beyond the ability to run GIMP. Exploitation is limited to the context of the logged-in user (S:U, scope unchanged), so attacker access is bounded by the victim's OS-level permissions. The mandatory user interaction (UI:R) is the primary limiting factor: automated exploitation without any user action is not possible. The specific configuration condition is simply having GIMP 3.2.2 installed and opening a crafted TIFF file - no non-default GIMP configuration is required.

The NVD CVSS 3.0 score of 7.8 (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) reflects high confidentiality, integrity, and availability impact but limits the attack vector to local execution requiring user interaction - the attacker must socially engineer a victim into opening a malicious TIFF file, either via a website or delivered file. This is a common and realistic delivery model for desktop image editor exploits. No EPSS data was provided, and the vulnerability is not listed in CISA KEV, indicating no confirmed active exploitation campaign at time of analysis. The absence of a tagged patched release (only a GitLab commit exists) means users on GIMP 3.2.2 remain exposed pending an official release. Real-world priority is moderate-to-high for organizations where GIMP is used to process untrusted image files, but limited by the mandatory user interaction requirement.

An attacker hosts a crafted TIFF file on a malicious web page or delivers it via email or messaging platform. When the victim opens the file in GIMP 3.2.2, the TIF parser copies attacker-controlled data into a heap buffer without length validation, triggering a heap overflow that redirects execution to attacker-supplied shellcode, running with the privileges of the current user account. No public exploit code has been identified at time of analysis, but the ZDI advisory publication and available fix commit increase the likelihood of future exploit development.