GIMP CVE-2025-10920

Exploitation requires the victim to open or preview a specially crafted ICNS (Apple Icon Image) file in GIMP; the flaw lies specifically in GIMP's ICNS file-format parser, which fails to validate user-supplied data before writing to an allocated buffer. The affected version per CPE is GIMP 3.0.4. Limiting factors: this is a local, client-side vector (AV:L) that cannot be triggered without user interaction (UI:R) - there is no automatic or network-service exposure - so an attacker must socially engineer the target into opening the malicious file, and code executes only with the privileges of the user running GIMP. A fixed version is available from the vendor.

This is a credible but not emergency-level priority. The CVSS 3.0 base score of 7.8 (HIGH) is driven by high confidentiality, integrity, and availability impact (C:H/I:H/A:H) under a local attack vector with low complexity and no privileges, offset only by required user interaction (UI:R). Note the vector is AV:L (local), NOT network - the 'remote attackers' language in the description reflects remote delivery of a malicious file that the victim must then open locally, which is the standard ZDI file-parsing pattern; this is consistent, not contradictory, but defenders should understand exploitation requires the victim to actively open or preview a crafted ICNS file, not a network-reachable service. No EPSS score was provided, so statistical exploitation probability cannot be weighed. The vulnerability is NOT in CISA KEV and no public exploit code is referenced, so there is no evidence of active exploitation. The ZDI advisory (ZDI-25-909) and GNOME GitLab merge request confirm a real, vendor-acknowledged memory-corruption flaw with a patch, but as a client-side desktop image-editor bug requiring user interaction, real-world risk is moderate: dangerous for targeted spear-phishing of individual users but unlikely to see widespread automated exploitation. No SSVC decision data was supplied in the input.

An attacker crafts a malicious ICNS image file and delivers it via phishing email, a download link, or a malicious web page, enticing a GIMP user to open or preview it. When GIMP parses the file, the lack of validation of user-supplied data triggers an out-of-bounds write past an allocated buffer, allowing the attacker to execute arbitrary code in the context of the GIMP process. No public proof-of-concept exploit is identified in the provided data.