PatchSiren cyber security CVE debrief
CVE-2026-63632 onnx CVE debrief
A low-severity vulnerability exists in Open Neural Network Exchange (ONNX) versions from 1.3.0 to 1.22.0. The onnx.version_converter.convert_version() function can perform an out-of-bounds read in Gemm_7_6::adapt_gemm_7_6() during an opset 7 to 6 downgrade when a Gemm node has input tensors with fewer than two dimensions. This issue can potentially cause a process crash. The vulnerability is fixed in version 1.22.0.
- Vendor
- onnx
- Product
- Unknown
- CVSS
- LOW 3.3
- CISA KEV
- Not listed in stored evidence
- Original CVE published
- 2026-08-18
- Original CVE updated
- 2026-09-18
- Advisory published
- 2026-08-18
- Advisory updated
- 2026-09-18
Who should care
Defenders responsible for machine learning systems using ONNX should assess exposure and prioritize verification and potential upgrades. This includes reviewing the current ONNX version, identifying systems that may be affected, and planning for upgrades or mitigations as needed. Additionally, defenders should monitor for potential crashes during opset downgrades and review compensating controls for exposed systems.
Why it matters
A low-severity vulnerability in Open Neural Network Exchange (ONNX) versions from 1.3.0 to 1.22.0 can cause a process crash during opset downgrades, requiring defenders to verify and potentially upgrade ONNX versions.
- Potential process crashes during opset downgrades
- Need for verification of ONNX version and potential upgrades
- Monitoring for crashes and implementing patches or workarounds
Technical summary
The onnx.version_converter.convert_version() function can perform an out-of-bounds read in Gemm_7_6::adapt_gemm_7_6() during an opset 7 to 6 downgrade when a Gemm node has input tensors with fewer than two dimensions, potentially causing a process crash. This issue arises from accessing B_shape[1], A_shape[0], or A_shape[1] without a rank check. The vulnerability is fixed in version 1.22.0, and defenders should prioritize verifying the ONNX version and upgrading if necessary, as well as monitoring for potential crashes during opset downgrades.
Defensive priority
Defenders should prioritize verifying the ONNX version and upgrading to 1.22.0 if necessary, as well as monitoring for potential crashes during opset downgrades.
Recommended defensive actions
- Verify the ONNX version and upgrade to 1.22.0 if necessary
- Monitor for potential crashes during opset downgrades
- Review and implement necessary patches or workarounds
- Confirm whether affected product deployments exist in managed environments and assign an owner for follow-up
- Review the supplied official advisory or CVE record to validate affected scope, severity, and vendor guidance
- Plan vendor-supported updates or mitigations through normal change control where exposure is confirmed
- Check relevant monitoring, detection, and logs for exposed assets that need extra review
Evidence notes
The CVE record and NVD entry provide details on the vulnerability, including its description, CVSS score, and affected versions. However, the corpus does not establish versions, exploitation, impact, or remediation beyond the provided information, which requires verification from the supplied official sources.
Sources and references
Verified primary and authoritative sources
-
CVE-2026-63632 CVE Program record
Publisher, destination, and source semantics verified
URL: https://www.cve.org/CVERecord?id=CVE-2026-63632
CVE Program - Official CVE Program record with source-provided CVE metadata.
-
CVE-2026-63632 NVD vulnerability detail
Publisher, destination, and source semantics verified
URL: https://nvd.nist.gov/vuln/detail/CVE-2026-63632
NIST National Vulnerability Database - Official NIST NVD detail page and source-specific vulnerability assessment.
Supplemental references
-
Source reference
Unverified legacy reference
URL: https://github.com/onnx/onnx/commit/e9c74f596eaa0250f89e52a54160a25bbcb25b66
-
Source reference
Unverified legacy reference
URL: https://github.com/onnx/onnx/pull/7880
-
Source reference
Unverified legacy reference
URL: https://github.com/onnx/onnx/releases/tag/v1.22.0
-
Source reference
Unverified legacy reference
URL: https://github.com/onnx/onnx/security/advisories/GHSA-p893-rvq9-2xf9
Methodology and review provenance
AI-assisted synthesis based on stored public vulnerability evidence. System validation, approval state, and publication status do not by themselves establish human review of this revision. PatchSiren helps prioritize defensive review and does not prove exposure or remediation on any system.