UTC

mlflow

91 published advisories; version 3.18.0 or later fixes all but 18 of them.

Registry
PyPI
Advisories
91
Latest fix
3.18.0
Exploited in the wild
1
Advisories with no fix yet

18 advisories have no fixed release in any version yet. What can be done meanwhile:

General guidance from ShieldPi, not a vendor fix

Check your projects

shieldpi agent-config checks agent and MCP packages from npm and PyPI, not mlflow. Any scanner that reads the OSV database finds it in your lockfiles, for example Google's open-source OSV-Scanner.

osv-scanner scan source -r .

Or check a lockfile against ShieldMap in your browser.

Provenance

Source: github.com/mlflow/mlflow, named in the package metadata (not verified)

10 of the last 10 releases of mlflow were published with build provenance: a signed record of the repository and commit each was built from, which the registry checks before it accepts the release.

  • 3.18.0
  • 3.17.2
  • 3.17.1
  • 3.17.0
  • 2.13.3
  • 2.11.5
  • 3.16.1
  • 3.16.0
  • 3.15.2
  • 3.15.1

Green: published with provenance. Grey: without.

OpenSSF Scorecard: 5.0 of 10 on 2026-11-09.

No dangerous CI workflows
0/10
CI tokens read-only by default
10/10
Build dependencies pinned
Not checked
Branch protection
3/10
Changes reviewed before merge
4/10
Signed releases
0/10
Actively maintained
10/10
Security policy
10/10
Full Scorecard →

Provenance from deps.dev (CC BY 4.0); Scorecard from OpenSSF (CDLA-Permissive-2.0). As of 2026-11-18.

Advisories
AdvisorySeverityAffectedFixed inPublished
Server-Side Request Forgery in the MLflow tracking server via client-supplied artifact locations and model version sourcesCVE-2026-118628High< 3.18.03.18.02026-11-13
Missing authorization on gateway secret creation in MLflow basic-auth with workspaces enabledGHSA-4449-4cjp-ffp5Medium>= 3.10.0, < 3.17.03.17.02026-10-28
MLflow vulnerable to remote code execution on unauthenticated servers via third-party scorer deserialization importing caller-placed artifacCVE-2026-116692Critical>= 3.12.0rc0, < 3.17.03.17.02026-10-27
MLflow CrewAI autologging records LLM provider API keys in logged trace spansCVE-2026-116691Medium>= 2.19.0, < 3.17.03.17.02026-10-27
MLflow vulnerable to arbitrary file read on unauthenticated servers via prompt model version sourceCVE-2026-116690High>= 3.10.0, < 3.17.03.17.02026-10-27
MLflow: SSRF on unauthenticated MLflow servers via third-party scorer kwargs in the scorer invoke, register, and GenAI evaluate endpointsCVE-2026-115886Medium>= 3.12.0rc0, < 3.17.03.17.02026-10-23
MLflow FileStore trace IDs allow path traversal and arbitrary file writesCVE-2026-53995High<= 3.17.13.17.22026-10-20
MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor — RCE via crafted model artifactGHSA-gqvg-gmmx-x4hmHigh>= 2.1.0, < 3.15.03.15.02026-09-01
MLflow: Unauthenticated full-read SSRF in webhook delivery: _validate_webhook_url bypassed via unvalidated HTTP redirects (and DNS rebindingCVE-2026-64849Exploited in the wildCritical>= 3.3.0, < 3.15.03.15.02026-08-17
MLflow AI Gateway permits SSRF through an unvalidated api_baseCVE-2026-71211High>= 3.13.0, <= 3.15.2No fix yet2026-08-05
LogInputs endpoint bypasses per-run UPDATE authorization in MLflow basic-authCVE-2026-69146Medium< 3.16.03.16.02026-08-04
CreateModelVersion source validation does not check READ permission on referenced run_idCVE-2026-69148Medium< 3.15.03.15.02026-08-04
MLflow: trace API endpoints lack proper authorization validatorsCVE-2026-8147High>= 2.14.0rc0, < 3.13.0rc03.13.0rc02026-07-02
MLflow: Deterministic sampling in dataset digest enables predictable collisionsCVE-2026-10803Low< 3.10.13.10.12026-06-04
MLflow: Environment variable injection in AI Gateway secrets enables server-side credential exfiltrationCVE-2026-4035Critical< 3.11.03.11.02026-06-03
MLflow: Any authenticated user can enumerate all gateway secrets, endpoints, and model definitionsCVE-2026-3198Medium< 3.11.0rc03.11.0rc02026-06-02
MLflow allows unauthorized access to multipart upload endpoints when the `--serve-artifacts` mode is enabledCVE-2026-2651Critical< 3.10.03.10.02026-05-26
MLflow authenticated users can enumerate any registered model versions due to lack of per-model permissions checksCVE-2026-2734Medium< 3.10.03.10.02026-05-21
MLflow: Improper Origin Validation in MLflow Assistant /ajax-api Endpoints Enables Browser-Mediated Local Command ExecutionCVE-2026-2611Critical= 3.9.03.10.02026-05-19
MLFlow Creates a Temporary File With Insecure PermissionsCVE-2026-4137High>= 1.26.0, < 3.11.03.11.02026-05-18
MLflow: unauthenticated access to certain FastAPI routesCVE-2026-2652High< 3.11.03.11.02026-05-15
MLflow allows an unauthenticated remote attacker to read arbitrary files from the server's filesystemCVE-2026-2614High< 3.10.03.10.02026-05-11
MLflow Has a Server-Side Request Forgery (SSRF) VulnerabilityCVE-2026-2393High< 3.9.03.9.02026-05-11
MLflow is vulnerable to an authorization bypass affecting the AJAX endpointCVE-2026-33866Medium<= 3.10.13.11.0rc02026-04-07
MLflow is vulnerable to Stored Cross-Site Scripting (XSS) caused by unsafe parsing of YAML-based MLmodel artifacts in its web interfaceCVE-2026-33865Medium<= 3.10.13.11.12026-04-07
mlflow: FastAPI job endpoints under `/ajax-api/3.0/jobs/*` are not protected by authentication or authorizationCVE-2026-0545Critical<= 3.10.1No fix yet2026-04-03
Mlflow: Command Injection when serving models with enable_mlserver=TrueCVE-2026-0596Critical< 3.9.03.9.02026-03-31
MLflow Command Injection vulnerabilityCVE-2025-15379Critical< 3.8.13.8.12026-03-30
MLFlow path traversal vulnerabilityCVE-2025-15036Critical< 3.9.0rc03.9.0rc02026-03-30
MLFlow allows Tracing + Assessments AccessCVE-2025-15381High<= 3.8.1No fix yet2026-03-27
Arbitrary file write via tar traversal in mlflowCVE-2025-15031High< 3.9.0rc03.9.0rc02026-03-19
MLflow has a command injection in mlflow/sagemaker/__init__.pyCVE-2025-14287High< 3.8.0rc03.8.0rc02026-03-16
MLflow Use of Default Password Authentication Bypass VulnerabilityCVE-2026-2635Critical< 3.8.0rc03.8.0rc02026-02-21
MLflow Tracking Server Artifact Handler Directory Traversal Remote Code Execution VulnerabilityCVE-2026-2033High< 3.8.0rc03.8.0rc02026-02-21
mlflow Creates of Temporary File in Directory with Insecure PermissionsCVE-2025-10279High< 3.4.0rc03.4.0rc02026-02-02
MLFlow is vulnerable to DNS rebinding attacks due to a lack of Origin header validationCVE-2025-14279High< 3.5.03.5.02026-01-12
MLflow Tracking Server Model Creation Directory Traversal Remote Code Execution VulnerabilityCVE-2025-11201High< 2.22.42.22.42025-10-29
MLflow Tracking Server Model Creation Directory Traversal Remote Code Execution VulnerabilityCVE-2025-11201High>= 3.0.0rc0, < 3.0.03.0.02025-10-29
MLflow Weak Password Requirements Authentication Bypass VulnerabilityCVE-2025-11200High< 2.22.0rc02.22.0rc02025-10-29
MLFlow SSRF via gateway_proxy_handlerCVE-2025-52967Medium< 2.22.22.22.22025-06-23
MLFlow SSRF via gateway_proxy_handlerCVE-2025-52967Medium>= 3.0.0rc0, < 3.1.03.1.02025-06-23
MLflow Uncontrolled Resource Consumption vulnerabilityCVE-2025-0453Medium<= 2.17.2No fix yet2025-03-20
MLflow has Weak Password RequirementsCVE-2025-1474Low< 2.19.02.19.02025-03-20
MLflow has a Local File Read/Path Traversal in dbfsCVE-2024-8859High< 2.17.0rc02.17.0rc02025-03-20
MLflow Cross-Site Request Forgery (CSRF) vulnerabilityCVE-2025-1473Medium>= 2.17.0, < 2.20.32.20.32025-03-20
MLflow Uncontrolled Resource Consumption vulnerabilityCVE-2024-6838Medium<= 2.13.2No fix yet2025-03-20
MLflow's excessive directory permissions allow local privilege escalationCVE-2024-27134High< 2.16.02.16.02024-11-25
Remote code execution in mlflowCVE-2024-0520Critical< 2.9.02.9.02024-06-06
Undefined Behavior in mlflowCVE-2024-3099Medium< 2.11.32.11.32024-06-06
Local File Inclusion in mlflowCVE-2024-2928High< 2.11.32.11.32024-06-06
MLFlow unsafe deserializationCVE-2024-37053High>= 1.1.0, <= 2.14.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37052High>= 1.1.0, <= 2.14.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37056High>= 1.23.0, <= 2.14.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37060High>= 1.27.0, <= 2.14.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37058High>= 2.5.0, <= 2.14.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37054High>= 0.9.0, <= 2.14.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37057High>= 2.0.0rc0, <= 2.14.1No fix yet2024-06-04
MLFlow improper input validationCVE-2024-37061High>= 1.11.0, <= 2.13.1No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37059High>= 0.5.0, <= 3.4.0No fix yet2024-06-04
MLFlow unsafe deserializationCVE-2024-37055High>= 1.24.0, <= 2.14.1No fix yet2024-06-04
MLflow allows low privilege users to delete any artifactCVE-2024-4263Medium< 2.10.12.10.12024-05-16
MLflow has a Local File Read/Path Traversal bypassCVE-2024-3848High>= 2.9.2, < 2.12.12.12.12024-05-16
mlflow vulnerable to Path TraversalCVE-2024-1560High<= 2.9.2No fix yet2024-04-16
mlflow vulnerable to Path TraversalCVE-2024-1593High<= 2.9.2No fix yet2024-04-16
mlflow Path Traversal vulnerabilityCVE-2024-1483High<= 2.9.22.12.12024-04-16
mlflow vulnerable to Path TraversalCVE-2024-3573Critical< 2.10.02.10.02024-04-16
mlflow vulnerable to Path TraversalCVE-2024-1558High<= 2.9.22.12.12024-04-16
mlflow vulnerable to Path TraversalCVE-2024-1594High<= 2.9.2No fix yet2024-04-16
MLFlow Cross-site Scripting vulnerability leads to client-side Remote Code ExecutionCVE-2024-27133Critical< 2.10.02.10.02024-02-24
Cross-site Scripting in MLFlowCVE-2024-27132Critical< 2.10.02.10.02024-02-24
MLflow Server-Side Request Forgery (SSRF)CVE-2023-6974Critical< 2.9.22.9.22023-12-20
MLflow Path Traversal VulnerabilityCVE-2023-6909High< 2.9.22.9.22023-12-20
MLFlow Path Traversal VulnerabilityCVE-2023-6975Critical< 2.9.22.9.22023-12-20
MLflow Local File Disclosure VulnerabilityCVE-2023-6977High< 2.9.22.9.22023-12-20
MLflow Path Traversal VulnerabilityCVE-2023-6976High< 2.9.22.9.22023-12-20
mlflow Command Injection vulnerabilityCVE-2023-6940High< 2.9.22.9.22023-12-19
Path traversal in MLflowCVE-2023-6831Critical< 2.9.22.9.22023-12-15
Path traversal in MLflowCVE-2023-6753High< 2.9.22.9.22023-12-13
Jinja2 template injection in mlflowCVE-2023-6709High< 2.9.22.9.22023-12-12
Cross-site Scripting (XSS) in MLflowCVE-2023-6568Medium< 2.9.02.9.02023-12-07
Information exposure in MLflowCVE-2023-43472High< 2.9.02.9.02023-12-05
MLflow authentication requirement bypass can allow a user to arbitrarily create an accountCVE-2023-6014Critical< 2.8.02.8.02023-11-16
Remote Code Execution due to Full Controled File Write in mlflowCVE-2023-6018Critical<= 2.8.12.9.22023-11-16
MLflow allowed arbitrary files to be PUT onto the serverCVE-2023-6015Critical< 2.8.12.8.12023-11-16
mlflow vulnerable to OS Command InjectionCVE-2023-4033High< 2.6.02.6.02023-08-01
MLflow Path Traversal vulnerabilityCVE-2023-3765Critical< 2.5.02.5.02023-07-19
mlflow Path Traversal vulnerabilityCVE-2023-2780Critical< 2.3.02.3.02023-05-17
mflow vulnerable to directory traversalCVE-2023-30172High< 2.0.0rc02.0.0rc02023-05-11
Remote file access vulnerability in `mlflow server` and `mlflow ui` CLIsGHSA-83fm-w79m-64r5Critical< 2.3.12.3.12023-05-01
Relative path traversal in mlflowCVE-2023-2356High< 2.3.12.3.12023-04-28
Remote file existence check vulnerability in `mlflow server` and `mlflow ui` CLIsCVE-2023-1176Medium<= 2.2.02.2.12023-03-24
mlflow is vulnerable to remote file access in `mlflow server` and `mlflow ui` CLIsCVE-2023-1177Critical<= 2.2.02.2.12023-03-24
Insecure Temporary File in mlflowCVE-2022-0736High< 1.23.11.23.12022-02-24

Sources: GitHub Advisory Database, reviewed and malware advisories (CC BY 4.0, https://github.com/advisories); advisories published by Cursor (https://github.com/cursor/cursor/security/advisories); advisories published by OpenClaw (https://github.com/openclaw/openclaw/security/advisories); advisories published by vLLM (https://github.com/vllm-project/vllm/security/advisories); advisories published by Open WebUI (https://github.com/open-webui/open-webui/security/advisories); advisories published by n8n (https://github.com/n8n-io/n8n/security/advisories); advisories published by Flowise (https://github.com/FlowiseAI/Flowise/security/advisories); advisories published by Langflow (https://github.com/langflow-ai/langflow/security/advisories); advisories published by the MCP Python SDK project (https://github.com/modelcontextprotocol/python-sdk/security/advisories); advisories published by MLflow (https://github.com/mlflow/mlflow/security/advisories); CISA Known Exploited Vulnerabilities catalog (CC0). Data as of 2026-11-22. ShieldMap describes how each problem works and how to fix it; it does not publish exploit code. Report a correction to [email protected].