CVE-2026-35444: Buffer Overflow in SDL_image

An attacker can exploit this vulnerability by crafting a malicious XCF image file. When SDLimage attempts to decode this file, the out-of-bounds read allows the attacker to potentially leak up to 762 bytes of heap memory. This leaked data, written into the output surface pixel data, could be observable in the rendered image, potentially revealing sensitive information. While direct code execution is unlikely, the information disclosure could be a stepping stone for further attacks, particularly in applications that rely on SDLimage to display user-provided images. The impact is amplified in scenarios where the application processes untrusted image data without proper sanitization.

1. Reserved2026-04-02
2. Published2026-04-06
3. Modified2026-04-08
4. EPSS updated2026-04-14

The primary mitigation is to upgrade to SDL_image version 2.0.6 or later, which contains the fix for this vulnerability. If upgrading is not immediately feasible, consider implementing input validation to restrict the types of image files processed by the application. Specifically, validate the XCF file format and its contents before attempting to decode it. WAFs or proxies can be configured to block requests containing potentially malicious XCF files, although this is a less robust solution. Monitor application logs for unusual memory access patterns or errors related to image decoding, which could indicate exploitation attempts. After upgrading, confirm the fix by attempting to load a known malicious XCF file and verifying that the out-of-bounds read no longer occurs.