# CVE-2026-112100: DeepPavlov code injection in MultiTaskPipelinePreprocessor

Source: https://junglewise.ai/threats/cve-2026-112100-a-vulnerability-was-determined-in-deeppavlov-up-to-1-7-0-this  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-10-19. Updated: 2026-10-19  
License: CC BY 4.0 with a link back

CVE-2026-112100 · Severity: medium · CVSS 5.5 · Published 2026-10-19

DeepPavlov is an open-source library for building chatbots and dialogue systems using deep learning. A code injection vulnerability in the MultiTaskPipelinePreprocessor component allows arbitrary Python code execution when loading a model configuration file, enabling attackers to compromise systems that run untrusted model packages.

Technologies: [DeepPavlov](https://junglewise.ai/threats/technologies/deeppavlov). Vendors: [DeepPavlov](https://junglewise.ai/threats/vendors/deeppavlov).

## Quick answers

**Is CVE-2026-112100 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-112100 being exploited in the wild.

**How severe is CVE-2026-112100?** Medium severity, with a CVSS score of 5.5, published 19 October 2026.

**Which products and versions are affected?** DeepPavlov up to 1.7.0.

**Is there a fix for CVE-2026-112100?** No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.

**How is CVE-2026-112100 attacked?** It is attacked over the network, in a software library.

**Am I affected by CVE-2026-112100?** Check whether you run DeepPavlov at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## Technical details

The MultiTaskPipelinePreprocessor.__init__ method uses Python's eval() function to evaluate the "preprocessor" parameter from a JSON configuration file without validation, allowing arbitrary code execution. The vulnerable parameter is attacker-controlled via the model config.json and reaches eval() without sanitization, making remote code execution possible if a victim loads a malicious model package via deeppavlov.build_model(), deeppavlov predict, or deeppavlov interact commands. The fix requires replacing eval() with safe class resolution through the component registry or an explicit allowlist.

## Affected products

- DeepPavlov DeepPavlov up to 1.7.0

## Timeline

- 2026-10-03: disclosed: Issue #1714 reported on GitHub
- 2026-10-19: advisory

## References

- [https://github.com/deeppavlov/DeepPavlov/](https://github.com/deeppavlov/DeepPavlov/)
- [https://github.com/deeppavlov/DeepPavlov/issues/1714](https://github.com/deeppavlov/DeepPavlov/issues/1714)

## Related threats

- [DeepPavlov code injection in config reference resolver](https://junglewise.ai/threats/cve-2026-112098-a-security-flaw-has-been-discovered-in-deeppavlov-up-to-1-7-0)
- [DeepPavlov path traversal in Model Download Handler](https://junglewise.ai/threats/cve-2026-109155-a-flaw-has-been-found-in-deeppavlov-up-to-1-7-0-this-impacts-the)