Skip to content
Heist/ feed
Back to feed
High

Kedro-Datasets: Remote code execution in experimental `PyTorchDataset` via unsafe `torch.load`

Type: advisorySources: OSV

Kedro-Datasets provides data connectors for Kedro. From version 5.0.0 until 9.5.0, kedro_datasets_experimental.pytorch.PyTorchDataset in kedro-datasets loads .pt model files with torch.load without enforcing weights_only=True, and user-supplied load_args are silently dropped. On PyTorch versions earlier than 2.6, a malicious pickle-backed model from an attacker-influenced shared registry, downloaded checkpoint, or partitioned external source can execute arbitrary code when a Kedro pipeline loads it. The issue affects only the opt-in kedro_datasets_experimental component and does not affect users who load only trusted files. This issue is fixed in version 9.5.0.

Products
Not specified
Affected versions
Not specified
Fixed versions
Not specified
CVEs
CVE-2026-62997
Confirmed fixed CVEs
Not specified
Updated in this feed
2 Nov 2026, 22:25 UTC · Revision 1

Source evidence

Kedro-Datasets: Remote code execution in experimental `PyTorchDataset` via unsafe `torch.load`

core-osv · Original source

Published 16 Sept 2026, 20:48 UTC

Kedro-Datasets provides data connectors for Kedro. From version 5.0.0 until 9.5.0, kedro_datasets_experimental.pytorch.PyTorchDataset in kedro-datasets loads .pt model files with torch.load without enforcing weights_only=True, and user-supplied load_args are silently dropped. On PyTorch versions earlier than 2.6, a malicious pickle-backed model from an attacker-influenced shared registry, downloaded checkpoint, or partitioned external source can execute arbitrary code when a Kedro pipeline loads it. The issue affects only the opt-in kedro_datasets_experimental component and does not affect users who load only trusted files. This issue is fixed in version 9.5.0.

osv:CVE-2026-62997
Complete source record
{
  "id": "6f667d43675f852818d3023f5d869d0c75da0bccf80cc4b2fb5c5014fe7aaedb",
  "url": "https://osv.dev/vulnerability/CVE-2026-62997",
  "method": "structured",
  "payload": {
    "kind": "advisory",
    "title": "Kedro-Datasets: Remote code execution in experimental `PyTorchDataset` via unsafe `torch.load`",
    "summary": "Kedro-Datasets provides data connectors for Kedro. From version 5.0.0 until 9.5.0, kedro_datasets_experimental.pytorch.PyTorchDataset in kedro-datasets loads .pt model files with torch.load without enforcing weights_only=True, and user-supplied load_args are silently dropped. On PyTorch versions earlier than 2.6, a malicious pickle-backed model from an attacker-influenced shared registry, downloaded checkpoint, or partitioned external source can execute arbitrary code when a Kedro pipeline loads it. The issue affects only the opt-in kedro_datasets_experimental component and does not affect users who load only trusted files. This issue is fixed in version 9.5.0.",
    "evidence": "osv:CVE-2026-62997",
    "products": [],
    "severity": "high",
    "structured": {
      "id": "CVE-2026-62997",
      "aliases": [
        "CVE-2026-62997",
        "GHSA-f9q4-h45w-jrrq"
      ],
      "metrics": [
        {
          "type": "CVSS_V4",
          "score": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N"
        }
      ],
      "affected": [
        {
          "ranges": [
            {
              "repo": "https://github.com/kedro-org/kedro-plugins",
              "type": "GIT",
              "events": [
                {
                  "introduced": "7b7f27c5bc036a6709690dac6f160dd09122f1ab"
                },
                {
                  "fixed": "fd3f7282ab870651888ebb34676bbd59261dd569"
                },
                {
                  "fixed": "4d6deb7a7ecba483d33fd43e2a6c063bcf22f413"
                }
              ],
              "database_specific": {
                "extracted_events": [
                  {
                    "range": [
                      {
                        "introduced": "5.0.0"
                      },
                      {
                        "fixed": "9.5.0"
                      }
                    ],
                    "source": "AFFECTED_FIELD",
                    "strategy": "StringRangeExpression"
                  },
                  {
                    "range": [
                      {
                        "fixed": "4d6deb7a7ecba483d33fd43e2a6c063bcf22f413"
                      }
                    ],
                    "source": "REFERENCES_COMMIT"
                  },
                  {
                    "range": [
                      {
                        "fixed": "fd3f7282ab870651888ebb34676bbd59261dd569"
                      }
                    ],
                    "source": "REFERENCES_TAG",
                    "original_tag": "kedro-datasets-9.5.0"
                  }
                ]
              }
            }
          ],
          "versions": [
            "kedro-datasets-9.4.0",
            "kedro-telemetry-0.8.0",
            "kedro-datasets-9.3.0",
            "kedro-datasets-9.2.0",
            "kedro-datasets-9.1.1",
            "kedro-datasets-9.1.0",
            "kedro-telemetry-0.7.0",
            "kedro-datasets-9.0.0",
            "kedro-telemetry-0.6.5",
            "kedro-datasets-8.1.0",
            "kedro-telemetry-0.6.4",
            "kedro-datasets-8.0.0",
            "kedro-airflow-0.10.0",
            "kedro-telemetry-0.6.3",
            "kedro-datasets-7.0.0",
            "kedro-datasets-6.0.0",
            "kedro-docker-0.6.2",
            "kedro-telemetry-0.6.2",
            "kedro-airflow-0.9.2",
            "kedro-datasets-5.1.0",
            "kedro-datasets-5.0.0"
          ],
          "database_specific": {
            "source": "https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2026-62997.json"
          }
        }
      ],
      "modified": "2026-10-28T02:30:24.138635852Z",
      "provider": "osv",
      "references": [
        "https://github.com/kedro-org/kedro-plugins/releases/tag/kedro-datasets-9.5.0",
        "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/62xxx/CVE-2026-62997.json",
        "https://github.com/kedro-org/kedro-plugins/security/advisories/GHSA-f9q4-h45w-jrrq",
        "https://nvd.nist.gov/vuln/detail/CVE-2026-62997",
        "https://github.com/kedro-org/kedro-plugins/issues/1431",
        "https://github.com/kedro-org/kedro-plugins/commit/4d6deb7a7ecba483d33fd43e2a6c063bcf22f413",
        "https://github.com/kedro-org/kedro-plugins/pull/1433"
      ],
      "provider_metadata": {
        "cwe_ids": [
          "CWE-502"
        ],
        "cna_assigner": "GitHub_M",
        "osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/62xxx/CVE-2026-62997.json"
      }
    },
    "related_cves": [
      {
        "value": "CVE-2026-62997",
        "evidence": "/id"
      }
    ],
    "security_relevant": true,
    "severity_evidence": "/severity/0/score"
  },
  "source_id": "core-osv",
  "document_id": 490547,
  "published_at": "2026-09-16T20:48:44.536+00:00"
}