ARTIFICE DAILY AI + CYBER SECURITY BRIEFING

Cybersecurity

Pillow update fixes code-execution flaws in image handling

Pillow’s 13.0.0 release closed several security bugs, including an EPS-handling issue that could lead to command execution through attacker-chosen filenames, alongside separate JPEG and TIFF weaknesses.

The Hacker News

Pillow’s maintainers released version 13.0.0 on 16 October and used it to close CVE-2026-112900. The flaw sat in EPS handling and only mattered on hosts that had Ghostscript installed.

The article said the attacker had to control the filename that Pillow passed along. In that setup, Ghostscript could treat the name as something to execute rather than just a path.

The bug covered Pillow from 2.6.0 to 12.3.0. The maintainers described it as moderate, while GitHub’s CVE entry gave it 9.2 out of 10.

The same release also fixed other issues in JPEG decoding and TIFF saving. The article said Anthropic used Claude to study open-source software security, and that Ada Logics manually validated CVE-2026-112403, which caused crashes rather than code execution.

A third flaw, CVE-2026-112393, had been present since 2013 and overflowed a buffer when compressed TIFF images were written to memory. The article advised upgrading to Pillow 13.0.0 and rebuilding containers, wheels or base images that still bundled older copies, while noting that Python 3.10 users would need to move to Python 3.11 or later.

Named in this story

Companies

OX Security
whose report drew attention to the flaws
GitHub
assigned the CVE record scores
Anthropic
used Claude to study open-source software security
Ada Logics
manually validated one JPEG bug

Governments and agencies

CISA
its analysts marked exploitation status in the CVE records

Products and systems

Pillow
the Python image library with the flaws
Ghostscript
the program Pillow called when opening EPS images
Claude
the model Anthropic used in that security study

How the source tells it

A mostly technical advisory that leans on security urgency, with risk framed through code-execution scenarios and severity scores.

  • urgency severity labels and direct upgrade guidance
  • threat framing attack paths are described through attacker-controlled filenames and command execution