{"id":"CVE-2026-93348","aliases":[],"summary":"Unsloth Zoo Code Injection via model_type in config.json","published_at":"2026-09-28T15:09:41.524Z","modified_at":"2026-10-01T22:23:00Z","products":[],"rating":{"severity":"high","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","attack_path":"internet_unauthenticated","user_interaction":true,"malicious":false,"red_hat":"Important","vendor_ratings":[{"source":"Red Hat","rating":"Important"}]},"fix_available":false,"kev":{"listed":false,"ransomware":false},"public_exploit":false,"exploits":[],"epss":{"score":0.004089999943971634,"percentile":0.3605799973011017,"date":"2026-11-20T00:00:00Z"},"priority":10.3,"record_count":2,"vendors":["Red Hat"],"details":"Unsloth Zoo versions 2025.9.9 before 2026.8.14, as implemented in Unsloth 2025.9.9 through 2026.8.19, contains a code injection vulnerability in the model-loading compile path where the get_transformers_model_type() function in hf_utils.py collects model_type values from nested model configurations without enforcing a character allowlist, allowing newlines and arbitrary Python source to survive normalization. Attackers can embed a newline in a nested model_type value within a malicious model's config.json to terminate the generated import statement and execute arbitrary Python code via exec() in unsloth_compile_transformers(), achieving remote code execution as the loading user when the model is loaded for training or inference.","affected_stack":[],"affected_other":[{"ecosystem":"CVE","package":"unslothai/unsloth","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"2025.9.9"},{"fixed":"2026.8.20"}]}],"fixed":["2026.8.20"],"provider":"cvelistv5","source":"CVE-2026-93348","relation":"self"},{"ecosystem":"CVE","package":"unslothai/unsloth-zoo","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"2025.9.9"},{"fixed":"2026.8.14"}]}],"fixed":["2026.8.14"],"provider":"cvelistv5","source":"CVE-2026-93348","relation":"self"}],"product_status":[{"vendor":"Red Hat","product":"Red Hat OpenShift AI (RHOAI)","cpe":"cpe:/a:redhat:openshift_ai","status":"affected","packages":["rhoai/odh-th06-cuda130-torch210-py312-rhel9","rhoai/odh-th06-rocm64-torch291-py312-rhel9"],"package_count":2,"fixed":[],"advisories":[],"note":"Affected","provider":"redhat-vex","source":"CVE-2026-93348"},{"vendor":"Red Hat","product":"Red Hat OpenShift AI (RHOAI)","cpe":"cpe:/a:redhat:openshift_ai","status":"not_affected","packages":["rhoai/odh-th06-cuda130-torch291-py312-rhel9"],"package_count":1,"fixed":[],"advisories":[],"note":"component_not_present","provider":"redhat-vex","source":"CVE-2026-93348"},{"vendor":"Red Hat","product":"Red Hat OpenShift AI 3.5","cpe":"cpe:/a:redhat:openshift_ai:3.5::el9","status":"fixed","packages":["rhoai/odh-th-torch-cuda-py312-rhel9","rhoai/odh-th-torch-rocm-py312-rhel9","rhoai/odh-training-cuda128-torch29-py312-rhel9","rhoai/odh-training-rocm64-torch29-py312-rhel9"],"package_count":4,"fixed":[],"advisories":["RHSA-2026:86891","RHSA-2026:95872"],"provider":"redhat-vex","source":"CVE-2026-93348"},{"vendor":"Red Hat","product":"Red Hat OpenShift AI 3.5","cpe":"cpe:/a:redhat:openshift_ai:3.5::el9","status":"not_affected","packages":["rhoai/odh-ai-gateway-operator-rhel9","rhoai/odh-ai-gateway-payload-processing-rhel9","rhoai/odh-automl-rhel9","rhoai/odh-autorag-rhel9","rhoai/odh-built-in-detector-rhel9","rhoai/odh-cli-rhel9","rhoai/odh-core-bff-rhel9","rhoai/odh-dashboard-operator-rhel9","rhoai/odh-dashboard-rhel9","rhoai/odh-data-science-pipelines-argo-argoexec-rhel9","rhoai/odh-data-science-pipelines-argo-workflowcontroller-rhel9","rhoai/odh-data-science-pipelines-operator-controller-rhel9","rhoai/odh-eval-hub-rhel9","rhoai/odh-feast-module-operator-rhel9","rhoai/odh-feast-operator-rhel9","rhoai/odh-feature-server-rhel9","rhoai/odh-fms-guardrails-orchestrator-rhel9","rhoai/odh-guardrails-detector-huggingface-runtime-rhel9","rhoai/odh-kf-notebook-controller-rhel9","rhoai/odh-kserve-agent-rhel9","rhoai/odh-kserve-autogluon-server-rhel9","rhoai/odh-kserve-controller-rhel9","rhoai/odh-kserve-llmisvc-controller-rhel9","rhoai/odh-kserve-localmodel-controller-rhel9","rhoai/odh-kserve-localmodelnode-agent-rhel9"],"package_count":119,"fixed":[],"advisories":[],"note":"vulnerable_code_not_present","provider":"redhat-vex","source":"CVE-2026-93348"}],"records":[{"provider":"cvelistv5","source_id":"CVE-2026-93348","relation":"self","url":"https://www.cve.org/CVERecord?id=CVE-2026-93348","license":"CC-BY-4.0","published_at":"2026-09-28T15:09:41.524Z","modified_at":"2026-09-30T14:43:10.011Z","summary":"Unsloth Zoo Code Injection via model_type in config.json","details":"Unsloth Zoo versions 2025.9.9 before 2026.8.14, as implemented in Unsloth 2025.9.9 through 2026.8.19, contains a code injection vulnerability in the model-loading compile path where the get_transformers_model_type() function in hf_utils.py collects model_type values from nested model configurations without enforcing a character allowlist, allowing newlines and arbitrary Python source to survive normalization. Attackers can embed a newline in a nested model_type value within a malicious model's config.json to terminate the generated import statement and execute arbitrary Python code via exec() in unsloth_compile_transformers(), achieving remote code execution as the loading user when the model is loaded for training or inference.","severity":[{"type":"CVSS_V4","score":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","source":"CVE-2026-93348"},{"type":"CVSS_V3","score":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N","source":"CVE-2026-93348"}],"affected":[{"ecosystem":"CVE","package":"unslothai/unsloth","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"2025.9.9"},{"fixed":"2026.8.20"}]}],"fixed":["2026.8.20"],"provider":"cvelistv5","source":"CVE-2026-93348","relation":"self"},{"ecosystem":"CVE","package":"unslothai/unsloth-zoo","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"2025.9.9"},{"fixed":"2026.8.14"}]}],"fixed":["2026.8.14"],"provider":"cvelistv5","source":"CVE-2026-93348","relation":"self"}],"product_status":[],"references":[{"type":"FIX","url":"https://github.com/unslothai/unsloth-zoo/pull/1083"},{"type":"FIX","url":"https://github.com/unslothai/unsloth-zoo/pull/1108"},{"type":"FIX","url":"https://github.com/unslothai/unsloth/commit/92ee0207a38ebc8af1a0e678aa92d7e7d901ea0d"},{"type":"WEB","url":"https://www.vulncheck.com/advisories/unsloth-zoo-code-injection-via-model-type-in-config-json"}]},{"provider":"redhat-vex","source_id":"CVE-2026-93348","relation":"self","url":"https://access.redhat.com/security/cve/CVE-2026-93348","license":"CC-BY-4.0","published_at":"2026-09-28T15:09:41.524Z","modified_at":"2026-10-01T22:23:00Z","summary":"unsloth-zoo: unsloth: Unsloth Zoo: Arbitrary code execution via crafted model configuration","details":"A flaw was found in Unsloth Zoo, which can lead to arbitrary code execution when loading a machine learning model. The vulnerability exists due to improper input sanitization of the model type parameter within configuration files during compilation. An attacker can exploit this issue by supplying a specially crafted model configuration containing embedded commands, resulting in the execution of arbitrary code with the privileges of the user running model training or inference.","severity":[{"type":"CVSS_V3","score":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","source":"CVE-2026-93348"},{"type":"Red Hat","score":"Important","source":"CVE-2026-93348"}],"affected":[],"product_status":[{"vendor":"Red Hat","product":"Red Hat OpenShift AI (RHOAI)","cpe":"cpe:/a:redhat:openshift_ai","status":"affected","packages":["rhoai/odh-th06-cuda130-torch210-py312-rhel9","rhoai/odh-th06-rocm64-torch291-py312-rhel9"],"package_count":2,"fixed":[],"advisories":[],"note":"Affected","provider":"redhat-vex","source":"CVE-2026-93348"},{"vendor":"Red Hat","product":"Red Hat OpenShift AI (RHOAI)","cpe":"cpe:/a:redhat:openshift_ai","status":"not_affected","packages":["rhoai/odh-th06-cuda130-torch291-py312-rhel9"],"package_count":1,"fixed":[],"advisories":[],"note":"component_not_present","provider":"redhat-vex","source":"CVE-2026-93348"},{"vendor":"Red Hat","product":"Red Hat OpenShift AI 3.5","cpe":"cpe:/a:redhat:openshift_ai:3.5::el9","status":"fixed","packages":["rhoai/odh-th-torch-cuda-py312-rhel9","rhoai/odh-th-torch-rocm-py312-rhel9","rhoai/odh-training-cuda128-torch29-py312-rhel9","rhoai/odh-training-rocm64-torch29-py312-rhel9"],"package_count":4,"fixed":[],"advisories":["RHSA-2026:86891","RHSA-2026:95872"],"provider":"redhat-vex","source":"CVE-2026-93348"},{"vendor":"Red Hat","product":"Red Hat OpenShift AI 3.5","cpe":"cpe:/a:redhat:openshift_ai:3.5::el9","status":"not_affected","packages":["rhoai/odh-ai-gateway-operator-rhel9","rhoai/odh-ai-gateway-payload-processing-rhel9","rhoai/odh-automl-rhel9","rhoai/odh-autorag-rhel9","rhoai/odh-built-in-detector-rhel9","rhoai/odh-cli-rhel9","rhoai/odh-core-bff-rhel9","rhoai/odh-dashboard-operator-rhel9","rhoai/odh-dashboard-rhel9","rhoai/odh-data-science-pipelines-argo-argoexec-rhel9","rhoai/odh-data-science-pipelines-argo-workflowcontroller-rhel9","rhoai/odh-data-science-pipelines-operator-controller-rhel9","rhoai/odh-eval-hub-rhel9","rhoai/odh-feast-module-operator-rhel9","rhoai/odh-feast-operator-rhel9","rhoai/odh-feature-server-rhel9","rhoai/odh-fms-guardrails-orchestrator-rhel9","rhoai/odh-guardrails-detector-huggingface-runtime-rhel9","rhoai/odh-kf-notebook-controller-rhel9","rhoai/odh-kserve-agent-rhel9","rhoai/odh-kserve-autogluon-server-rhel9","rhoai/odh-kserve-controller-rhel9","rhoai/odh-kserve-llmisvc-controller-rhel9","rhoai/odh-kserve-localmodel-controller-rhel9","rhoai/odh-kserve-localmodelnode-agent-rhel9"],"package_count":119,"fixed":[],"advisories":[],"note":"vulnerable_code_not_present","provider":"redhat-vex","source":"CVE-2026-93348"}],"references":[{"type":"WEB","url":"https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-93348.json"},{"type":"WEB","url":"https://access.redhat.com/security/cve/CVE-2026-93348"},{"type":"WEB","url":"https://bugzilla.redhat.com/show_bug.cgi?id=2542526"},{"type":"WEB","url":"https://www.cve.org/CVERecord?id=CVE-2026-93348"},{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-93348"},{"type":"WEB","url":"https://github.com/unslothai/unsloth-zoo/pull/1083"},{"type":"WEB","url":"https://github.com/unslothai/unsloth-zoo/pull/1108"},{"type":"WEB","url":"https://github.com/unslothai/unsloth/commit/92ee0207a38ebc8af1a0e678aa92d7e7d901ea0d"},{"type":"WEB","url":"https://www.vulncheck.com/advisories/unsloth-zoo-code-injection-via-model-type-in-config-json"}]}],"references":[{"type":"FIX","url":"https://github.com/unslothai/unsloth-zoo/pull/1083"},{"type":"FIX","url":"https://github.com/unslothai/unsloth-zoo/pull/1108"},{"type":"FIX","url":"https://github.com/unslothai/unsloth/commit/92ee0207a38ebc8af1a0e678aa92d7e7d901ea0d"},{"type":"WEB","url":"https://www.vulncheck.com/advisories/unsloth-zoo-code-injection-via-model-type-in-config-json"},{"type":"WEB","url":"https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-93348.json"},{"type":"WEB","url":"https://access.redhat.com/security/cve/CVE-2026-93348"},{"type":"WEB","url":"https://bugzilla.redhat.com/show_bug.cgi?id=2542526"},{"type":"WEB","url":"https://www.cve.org/CVERecord?id=CVE-2026-93348"},{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-93348"}],"sources":[{"provider":"cvelistv5","source_id":"CVE-2026-93348","url":"https://www.cve.org/CVERecord?id=CVE-2026-93348","license":"CC-BY-4.0","modified_at":"2026-09-30T14:43:10.011Z","relation":"self"},{"provider":"redhat-vex","source_id":"CVE-2026-93348","url":"https://access.redhat.com/security/cve/CVE-2026-93348","license":"CC-BY-4.0","modified_at":"2026-10-01T22:23:00Z","relation":"self"}]}
