thumbnail Python package: OS command injection in generate_thumbnail

Summary

The thumbnail Python package through version 1.5 contains an OS command injection vulnerability in generate_thumbnail(). Shell metacharacters in file paths or option values are passed to os.system() calls for ffmpeg, convert or unoconv, so an attacker who controls a file name, such as an uploaded file, can run arbitrary commands with the privileges of the Python process. A public proof of concept exists, and CISA SSVC rates exploitation as PoC with automatable exploitation and total technical impact.

Affected

• thumbnail Python package, versions 0 to 1.5 (including 1.5)

Fixed in

• Unknown. No fixed release has been published. A proposed fix (GitHub pull request #9) is open and not merged.

Workaround

• Do not pass user-supplied file names or paths to generate_thumbnail(); rename uploads to server-generated names before processing.
• Reject input and option values containing shell metacharacters.
• Run thumbnail generation in an isolated container or sandbox with a non-privileged user, no network access and a read-only filesystem except a temporary directory.
• Remove the package if it is not needed, or replace it with your own subprocess-based code.

Exploitation

• Public proof of concept: Yes (reproduction in the pull request; not reproduced here).
• Known exploitation in the wild: None reported.
• CISA KEV: Not listed according to available sources (unverified directly).
• CISA SSVC: Exploitation poc, automatable yes, technical impact total.

Our recommendation

• Check whether any internal or customer-facing Python application imports the thumbnail package (for example with pip freeze or a software inventory).
• Treat internet-facing upload features that use it as urgent and apply the workarounds above.
• Monitor the upstream repository and PyPI for a fixed release and update promptly once available.
• Review logs of upload and thumbnail processing for unusual file names or unexpected child processes.

Public exploits iAttack code is publicly available (PoC on GitHub, Metasploit or Exploit-DB).

PoC-in-GitHub, Metasploit and Exploit-DB list nothing for this CVE.

Research sources

• VulnCheck advisory www.vulncheck.com
• GitHub pull request #9 (proposed fix, not merged) github.com
• Vulnerable source file thumbnail.py github.com
• VulDB entry vuldb.com
• Mend via Vulners vulners.com
• Strix CVE summary www.strix.ai

Researched and written up in English and German by an AI (Claude by Anthropic) based on NVD, CISA, EPSS and the sources listed. The vendor information is authoritative. Researched on 31 Oct 2026, 02:03.