{"id":"CVE-2026-115920","aliases":["GHSA-6pwm-x34p-fvj2"],"title":"ONNX: Out-of-bounds read in ConvTranspose shape inference via undersized weight rank","summary":"ONNX: Out-of-bounds read in ConvTranspose shape inference via undersized weight rank","severity":"medium","cvss":6.2,"cwe":["CWE-125"],"vendor":"onnx","product":"onnx","ecosystem":"pip","affected":["onnx >= 1.2.1, <= 1.21.0"],"patched":["onnx 1.22.0"],"published":"2026-10-28","updated":"2026-10-28","sourceUpdated":"2026-10-28T21:16:43Z","source":"GHSA","sourceUrl":"https://github.com/advisories/GHSA-6pwm-x34p-fvj2","references":[{"url":"https://github.com/onnx/onnx/security/advisories/GHSA-6pwm-x34p-fvj2"},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-115920"},{"url":"https://github.com/onnx/onnx/pull/7997"},{"url":"https://github.com/onnx/onnx/commit/2ec288908952eea9dcf48d4b197915e1f72c6abf"},{"url":"https://github.com/onnx/onnx/releases/tag/v1.22.0"},{"url":"https://github.com/advisories/GHSA-6pwm-x34p-fvj2"}],"tags":["ghsa","pip","cve.org"],"epss":0.00178,"epssPercentile":0.07376,"ssvc":{"exploitation":"none","automatable":"no","technicalImpact":"partial","timestamp":"2026-10-26T19:30:01.679717Z"},"ingestedAt":"2026-10-26T17:33:33.052Z","slug":"CVE-2026-115920","body":"## Overview\n\n## Summary\n\nONNX ConvTranspose shape inference performed an out-of-bounds read when the weight tensor had fewer than three dimensions and the model did not provide a `kernel_shape` attribute.\n\nThe inference function derived an empty `effective_kernel_shape` vector from the undersized weight tensor, then indexed that vector once for each spatial dimension of the input tensor. A crafted model passed `onnx.checker.check_model()` and caused SIGSEGV when passed to `onnx.shape_inference.infer_shapes()`.\n\n## Technical details\n\nFor an input shape such as `[1, 1, 5]`, shape inference computes one spatial input dimension. A weight shape such as `[1, 3]` has no spatial dimensions, so deriving `kernel_shape` from dimensions 2 onward produces an empty vector. The output loop nevertheless accesses `effective_kernel_shape[0]`.\n\nThe defect affected both the older `convTransposeShapeInference_opset11()` path and the then-current ConvTranspose inference implementation.\n\n## Impact\n\nAn application that performs shape inference on an untrusted model can be terminated, causing denial of service. The observed result was a deterministic process crash. No confidentiality or integrity impact was demonstrated.\n\n## Affected versions\n\n- Affected: ONNX 1.2.1 through 1.21.0\n- Patched: ONNX 1.22.0\n\n## Remediation\n\nUpgrade to ONNX 1.22.0 or later. The fix rejects input and weight tensors with fewer than three dimensions before deriving the spatial kernel shape.\n\n- Fix: [onnx/onnx#7997](https://github.com/onnx/onnx/pull/7997)\n- Commit: [`2ec288908`](https://github.com/onnx/onnx/commit/2ec288908952eea9dcf48d4b197915e1f72c6abf)\n\nThe original malformed model was re-tested against the published ONNX 1.23.2 wheel and correctly raises `InferenceError`.\n\n## Severity\n\nThe vulnerable ONNX component processes the malicious model through a local library API or file boundary and is not itself bound to a network stack. Applications that expose ONNX model processing over a network should calculate an Environmental score with Modified Attack Vector set to Network.\n\n## Affected packages\n\n- `onnx >= 1.2.1, <= 1.21.0`\n\n## Remediation\n\nUpgrade to a patched release:\n\n- `onnx 1.22.0`","depth":"sunlit","depthScore":34,"depthScoreParts":{"impact":34.1,"likelihood":0,"exploitation":0,"ransomware":0},"changes":[]}