# CVE-2026-90777

## Summary

- **CVE ID:** CVE-2026-90777
- **Severity:** HIGH
- **CVSS Score:** 8.8 (CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N)
- **CWE:** CWE-502
- **Published:** Sep 13, 2026
- **Last Modified:** Sep 14, 2026

## Description

ESPnet before 202609 deserializes pretrained model checkpoints using torch.load with weights_only=False, allowing arbitrary code execution from attacker-supplied files. Attackers can craft malicious checkpoint files that execute code during deserialization when loaded through the initialization or fine-tuning path.

## Affected Products

- espnet — espnet (0)
- espnet — espnet (202609)

## References

- [CNA](https://github.com/espnet/espnet/security/advisories/GHSA-64f6-3gqc-r926)
- [CNA](https://github.com/espnet/espnet/commit/91ca045fc179f29bc7b7fe05cd852aea65310003)
- [CNA](https://github.com/espnet/espnet/blob/v.202511/espnet2/torch_utils/load_pretrained_model.py#L99)
- [CNA](https://github.com/espnet/espnet/releases/tag/v.202609)
- [CNA](https://github.com/espnet/espnet)
- [CNA](https://www.vulncheck.com/advisories/espnet-before-202609-remote-code-execution-via-unsafe-deserialization)

## Exploitation Prediction (EPSS)

- **EPSS Score:** 0.51%
- **EPSS Percentile:** 42.1

---
_Exported from OnDuty AI Vulnerability Intelligence on 2026-09-18._