CVE-2026-108954: k2-fsa sherpa-onnx Offline Transducer Greedy Search Decoder offline-transducer-greedy-search-decoder.cc Decode buffer overflow

Published Oct 13, 2026
·
Updated

A vulnerability was found in k2-fsa sherpa-onnx up to 1.13.8. This issue affects the function OfflineTransducerGreedySearchDecoder::Decode of the file sherpa-onnx/csrc/offline-transducer-greedy-search-decoder.cc of the component Offline Transducer Greedy Search Decoder. The manipulation of the argument vocabsize results in buffer overflow. The attack can be launched remotely. The exploit has been made public and could be used. If only official model packs are used and they always keep metadata consistent with the joiner, the bug may not be reachable in normal supported usage. However, from a security and robustness standpoint, the code still contains the vulnerability because it trusts unvalidated model metadata and can be triggered by a malformed or tampered model pack.

Affected Software

1 affected component
k2-fsa sherpa-onnx<=1.13.8

Event History

Oct 13, 2026
CVE Published
via MITRE·01:30 AM
Data Sourced
via MITRE·01:30 AM
DescriptionSeverityWeakness
Data Sourced
via NVD·02:16 AM
DescriptionSeverityWeakness

Frequently Asked Questions

1

Who is most likely to be exposed to this issue?

Deployments that load malformed or tampered sherpa-onnx model packs are exposed, because the decoder trusts model metadata. Normal supported use may not reach the flaw when only official model packs are used and their metadata remains consistent with the joiner.

2

What does an attacker need to exploit the vulnerability?

An attacker needs to cause processing of a model pack with a manipulated vocab_size value. The issue can be launched remotely, requires no privileges, and requires user interaction according to the supplied severity vector.

3

What can be done if updating is not immediately possible?

Restrict model packs to official, trusted sources and ensure their metadata is consistent with the joiner. Do not load unvalidated, modified, or externally supplied model packs.

4

How can we determine whether our deployment may be affected?

Check whether sherpa-onnx is used at version 1.13.8 or earlier and whether the application accepts or loads model packs that could be tampered with or malformed. Review the model-loading path for control over metadata, particularly vocab_size.

Contact

SecAlerts Pty Ltd.
132 Wickham Terrace
Fortitude Valley,
QLD 4006, Australia
info@secalerts.co
By using SecAlerts services, you agree to our services end-user license agreement. This website is safeguarded by reCAPTCHA and governed by the Google Privacy Policy and Terms of Service. All names, logos, and brands of products are owned by their respective owners, and any usage of these names, logos, and brands for identification purposes only does not imply endorsement. If you possess any content that requires removal, please get in touch with us.
© 2026 SecAlerts Pty Ltd.
ABN: 70 645 966 203, ACN: 645 966 203