Pillow EUVD-2026-104638

No special configuration is required - any application using Pillow versions 5.4.0 through pre-13.0.0 that calls Image.open() on an attacker-supplied GIF is exploitable against default behavior, because frame zero (including comment extensions) is parsed automatically on open. The specific trigger is a GIF containing many comment extensions / comment sub-blocks, which drives quadratic copying in GifImageFile._seek(). Limiting factors: the attacker must be able to get a malicious GIF processed by the target (e.g. an upload, avatar, or thumbnail pipeline); impact is bounded to availability/CPU (VA:L) with no code execution or data exposure; and environments that only process trusted images are not practically exposed.

This is a denial-of-service issue (CWE-407, inefficient algorithmic complexity) with a vendor-provided CVSS 4.0 score of 6.9. The vector AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N is internally consistent with the description: remote, low-complexity, unauthenticated, no user interaction, and impact limited to availability only (VA:L - no confidentiality or integrity effect). No EPSS score was supplied, the CVE is NOT on CISA KEV, and no public exploit code is referenced, so the exploitation status is: no public exploit identified at time of analysis. The signals are coherent rather than conflicting - this is a genuine, easily-triggered DoS against any service that opens untrusted GIFs with Pillow, but the impact ceiling is low (CPU exhaustion / service slowdown, not RCE or data loss), which is why the score lands in the medium range rather than high/critical. Priority should be driven by exposure: high for services that accept user-uploaded images, low for environments that only process trusted files. Note: the input Tag 'Information Disclosure' contradicts the CVSS (VC:N/VI:N) and the description - this is purely an availability flaw, not information disclosure.

An attacker uploads a small-dimension GIF crafted with a large number of comment extension sub-blocks to any application that opens images with Pillow (e.g. an image-upload or thumbnailing endpoint). Because frame zero is parsed during Image.open() and comment bytes are copied quadratically in GifImageFile._seek(), a tiny file forces disproportionate CPU consumption and degrades or denies service before the app can even inspect the image. No public exploit code is referenced, but the trigger is trivial to reproduce from the description.