{"id":"CVE-2026-76395","aliases":["EUVD-2026-62960","GHSA-wmwx-pqh3-xv86"],"status":"published","title":"Remote Code Execution (RCE) through Deserialization of Untrusted Data in the Model Loading REST API in Splunk AI Toolkit","description":"In Splunk AI Toolkit versions below 6.0.0, a user who holds the \"power\" Splunk role could execute arbitrary code on the Splunk server by loading a model file containing crafted sparse matrix data. The deserialization of untrusted data is possible because a model codec in Splunk AI Toolkit deserializes sparse matrix data without guarding against embedded pickle content. For more information see Troubleshoot the Splunk Machine Learning Toolkit (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/machine-learning-toolkit-user-guide/5.5.0/troubleshooting-mltk/troubleshoot-the-splunk-machine-learning-toolkit) in the Splunk documentation.","severity":{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-11","score":0.00652,"percentile":0.51702,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":false},"cwes":["CWE-502"],"published":"2026-08-19T21:35:12Z","modified":"2026-08-26T20:18:01Z","products":[{"vendor":"splunk","product":"ai_toolkit"},{"vendor":"splunk","product":"splunk ai toolkit"}],"references":[{"url":"https://advisory.splunk.com/advisories/SVD-2026-0808","tags":["Vendor Advisory"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-76395"},{"url":"https://github.com/advisories/GHSA-wmwx-pqh3-xv86"}],"sources":{"cve":[{"id":"CVE-2026-76395","title":"Remote Code Execution (RCE) through Deserialization of Untrusted Data in the Model Loading REST API in Splunk AI Toolkit","description":"In Splunk AI Toolkit versions below 6.0.0, a user who holds the \"power\" Splunk role could execute arbitrary code on the Splunk server by loading a model file containing crafted sparse matrix data. The deserialization of untrusted data is possible because a model codec in Splunk AI Toolkit deserializes sparse matrix data without guarding against embedded pickle content. For more information see Troubleshoot the Splunk Machine Learning Toolkit (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/machine-learning-toolkit-user-guide/5.5.0/troubleshooting-mltk/troubleshoot-the-splunk-machine-learning-toolkit) in the Splunk documentation.","status":"published","published":"2026-08-19T21:35:12.154Z","modified":"2026-08-26T19:51:54.013Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","level":"high","assessor":"cna"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T13:46:32.755125Z"}],"euvd":[{"id":"EUVD-2026-62960","description":"In Splunk AI Toolkit versions below 6.0.0, a user who holds the \"power\" Splunk role could execute arbitrary code on the Splunk server by loading a model file containing crafted sparse matrix data. The deserialization of untrusted data is possible because a model codec in Splunk AI Toolkit deserializes sparse matrix data without guarding against embedded pickle content. For more information see Troubleshoot the Splunk Machine Learning Toolkit (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/machine-learning-toolkit-user-guide/5.5.0/troubleshooting-mltk/troubleshoot-the-splunk-machine-learning-toolkit) in the Splunk documentation.","status":"published","published":"2026-08-19T21:35:12Z","modified":"2026-08-26T19:51:54Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","level":"high"}],"updated_at":"2026-11-02T18:01:29.296861Z"}],"ghsa":[{"id":"GHSA-wmwx-pqh3-xv86","title":"In Splunk AI Toolkit versions below 6.0.0, a user who holds the \"power\" Splunk role could execute...","description":"In Splunk AI Toolkit versions below 6.0.0, a user who holds the \"power\" Splunk role could execute arbitrary code on the Splunk server by loading a model file containing crafted sparse matrix data. The deserialization of untrusted data is possible because a model codec in Splunk AI Toolkit deserializes sparse matrix data without guarding against embedded pickle content. For more information see Troubleshoot the Splunk Machine Learning Toolkit (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/machine-learning-toolkit-user-guide/5.5.0/troubleshooting-mltk/troubleshoot-the-splunk-machine-learning-toolkit) in the Splunk documentation.","status":"published","published":"2026-08-20T00:35:06Z","modified":"2026-08-20T00:35:11Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","level":"high"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T15:10:45.33933Z"}],"nvd":[{"id":"CVE-2026-76395","description":"In Splunk AI Toolkit versions below 6.0.0, a user who holds the \"power\" Splunk role could execute arbitrary code on the Splunk server by loading a model file containing crafted sparse matrix data. The deserialization of untrusted data is possible because a model codec in Splunk AI Toolkit deserializes sparse matrix data without guarding against embedded pickle content. For more information see Troubleshoot the Splunk Machine Learning Toolkit (https://help.splunk.com/en/splunk-cloud-platform/apply-machine-learning/machine-learning-toolkit-user-guide/5.5.0/troubleshooting-mltk/troubleshoot-the-splunk-machine-learning-toolkit) in the Splunk documentation.","status":"published","published":"2026-08-19T22:17:26.023Z","modified":"2026-08-26T20:18:01.2Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","level":"high","assessor":"secondary"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T14:12:39.962615Z"}]}}