libheif CVE-2026-32740

Requires the victim's libheif build to handle a HEIF or AVIF file containing a grid-derived image with a 1×4 tile layout where the tiles have odd pixel height - the description states the overflow is triggered 'during normal image decoding with default build configuration', so no non-default compile flags or runtime options are needed. The only limiting factor is CVSS UI:R: the file must be opened, previewed, or auto-thumbnailed by an application linking libheif ≤ 1.21.2; no authentication, network listener, or special privileges are needed on the target. Platform is anywhere libheif runs (Linux, macOS, Windows, Android, iOS via downstream ports).

The CVSS 3.1 vector AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H scores 8.8: network-reachable (any file delivery channel), low complexity, no privileges, but user interaction required to open/preview the file. SSVC-style read: technical impact is total (heap write with full content control), automatable is plausible since the trigger is just a malformed file served via web/email/messenger, but exploitation status is currently 'none' - no KEV listing and no public exploit identified at time of analysis. EPSS data was not provided in the input. The real-world risk is meaningfully elevated above a typical heap overflow because libheif is silently invoked by OS thumbnailers and chat clients, turning 'UI:R' into little more than a file arriving in a folder; treat as a high-priority patch even without confirmed in-the-wild use.

An attacker emails or messages a victim a HEIC/AVIF photo whose grid box describes a 1×4 layout with odd-height tiles; when the victim's mail client, browser, or file manager auto-generates a thumbnail, libheif's grid compositor writes 64 attacker-chosen bytes past a chroma-plane heap chunk, corrupting adjacent allocator metadata or object pointers to pivot toward code execution in the rendering process. The CVSS UI:R requirement is satisfied by routine preview behavior rather than an explicit double-click. No public exploit identified at time of analysis, but the fully-controlled overflow content makes weaponization tractable.