VYPR
Vypr IntelligenceAI-generatedOct 19, 2026· 25 CVEs

Pillow: 25 Vulnerabilities Disclosed Together, Ranging from Critical to Medium

A batch of 25 vulnerabilities, ranging from Medium to Critical, was disclosed in Pillow on October 19, 2026, affecting versions prior to 13.0.0.

Key findings

  • 25 vulnerabilities disclosed in Pillow on October 19, 2026, ranging from Medium to Critical severity.
  • Multiple memory corruption issues including use-after-free, buffer overflows, and reading uninitialized memory.
  • Denial-of-service vulnerabilities stemming from inefficient parsing of image formats like XBM, PDF, and GIF.
  • Critical flaws in JPEG, FlashPix, and EPS handling, posing significant security risks.
  • All disclosed vulnerabilities are fixed in Pillow version 13.0.0.
  • Affected versions are prior to 13.0.0.

On October 19, 2026, a significant batch of 25 vulnerabilities was disclosed in Pillow, the Python Imaging Library. These vulnerabilities, disclosed within a one-hour window, span a range of severity levels from Medium to Critical, with many High-severity flaws identified. The disclosures highlight potential risks including denial-of-service, memory corruption, information disclosure, and resource exhaustion, primarily stemming from vulnerabilities in how Pillow handles various image formats and external libraries like libtiff and Ghostscript.

Several vulnerabilities are related to memory safety and corruption. CVE-2026-112898 and CVE-2026-112897 describe issues where Pillow can disclose uninitialized stack memory or return uninitialized memory as pixels, respectively, when handling specific image formats or encountering errors during decoding. CVE-2026-112896 points to a vulnerability in the TIFF encoder that can read beyond a buffer when metadata is inconsistent, potentially leading to memory corruption. Similarly, CVE-2026-112893 details a use-after-free vulnerability during FreeType rendering, and CVE-2026-112402 describes a double-free vulnerability in memory management during affine transformations. CVE-2026-112396 also involves a heap use-after-free when libtiff encounters errors during TIFF encoding.

Denial-of-service (DoS) vulnerabilities are also present in this batch. CVE-2026-112901 describes a regular-expression denial-of-service in the XBM header parser, where crafted XBM files can cause excessive backtracking. CVE-2026-112895 and CVE-2026-112406 highlight vulnerabilities in PDF parsing where Pillow can exhaust resources or enter an infinite loop when processing malformed PDF files due to improper handling of page-tree references and object identifiers. CVE-2026-112894 also contributes to DoS by taking quadratic time to reserve PDF object identifiers when saving multi-frame images. CVE-2026-112401 details a quadratic copying vulnerability when parsing GIF comments, which can be exploited with specially crafted GIFs.

Critical vulnerabilities in this batch include CVE-2026-112900, an argument injection flaw when invoking Ghostscript for EPS input, allowing for command execution. CVE-2026-112403 describes a JPEG decoder vulnerability where Pillow can write beyond a buffer sized for a draft scale, potentially leading to memory corruption. Additionally, CVE-2026-112398 presents a FlashPix JPEG-tile decoder vulnerability that can write decoded scanlines beyond allocated buffers, and CVE-2026-112393 details a heap buffer overflow and stale-heap disclosure when saving compressed TIFF data.

Other notable vulnerabilities include CVE-2026-112892, where the MSP RLE decoder can expand a small file into an unbounded buffer, and CVE-2026-112400, which can retain a rejected GIF frame size after a decompression-bomb exception, potentially leading to issues if processing continues. CVE-2026-112392 involves dereferencing invalid image storage for zero-size images, and CVE-2026-112395 can read past a tile buffer when copying four-byte pixels in tiled-TIFF decoding. CVE-2026-112404 and CVE-2026-112394 describe issues with raw decoders reading past input buffers or end of buffers due to discrepancies in how Pillow and underlying libraries interpret image dimensions and data. CVE-2026-112397 involves memory access before a row buffer for zero-width images in morphology operations.

All disclosed vulnerabilities affect versions prior to Pillow 13.0.0. Users are strongly advised to update to Pillow 13.0.0 or later to mitigate these risks. The wide range of issues underscores the importance of keeping image processing libraries up-to-date, as vulnerabilities can arise from complex interactions with various file formats and external dependencies.

This extensive batch of vulnerabilities highlights the critical need for robust security practices when handling image processing in applications. Developers relying on Pillow should prioritize updating to the patched version to protect against potential exploits that could lead to denial-of-service, data corruption, or information disclosure.

The vulnerabilities were disclosed on October 19, 2026, and affect Pillow versions prior to 13.0.0. The patched version is 13.0.0.

Key findings include:

  • 25 vulnerabilities disclosed in Pillow on October 19, 2026, ranging from Medium to Critical severity.
  • Multiple memory corruption issues including use-after-free, buffer overflows, and reading uninitialized memory.
  • Denial-of-service vulnerabilities stemming from inefficient parsing of image formats like XBM, PDF, and GIF.
  • Critical flaws in JPEG, FlashPix, and EPS handling, posing significant security risks.
  • All disclosed vulnerabilities are fixed in Pillow version 13.0.0.
  • Affected versions are prior to 13.0.0.

A stylized Python logo is shown with several image file icons (JPEG, TIFF, PDF, GIF) being processed, with some icons glitching or showing error symbols, indicating vulnerabilities. The background is a dark, abstract representation of code.

AI-written article. Grounded in 25 CVE records listed below.