CVE-2024-37061

MLflow CVE-2024-37061 affects MLflow platforms running version 1.11.0 or newer. A maliciously crafted MLproject can trigger remote code execution on an end user’s system when run. Multiple connected sources corroborate RCE related to MLflow projects input handling, including descriptions and advi...

CVE-2024-37059

MLflow contains a deserialization vulnerability affecting versions 0.5.0 and later in the PyTorch integration (mlflow/pytorch/init .py _load_model), where untrusted model data can trigger arbitrary code execution when a PyFunc model is loaded. Reports from Veracode describe attacker-controlled pi...

CVE-2023-1177

CVE-2023-1177 is a path traversal / local-file-inclusion vulnerability in mlflow/mlflow prior to 2.2.1. Affects the MLflow server endpoints handling model-versions and artifacts (e.g., GET /model-versions/get-artifact) where a crafted path can disclose files outside the intended directory. Concre...

CVE-2026-64849

MLflow versions before 3.15.0 expose an unauthenticated full-read SSRF: _validate_webhook_url() in mlflow/utils/validation.py validates only the original URL while mlflow/webhooks/delivery.py follows HTTP redirects without IP pinning, letting attackers reach internal/cloud-metadata endpoints. CIR...

CVE-2024-3848

CVE-2024-3848 affects mlflow/mlflow up to version 2.11.0. The path traversal stems from improper handling of the fragment component in artifact URLs, where a ‘#’ can insert a path into the URL fragment and bypass validation, allowing an attacker to read arbitrary files by mapping the URL to a fil...

CVE-2022-0736

CVE-2022-0736 affects mlflow/mlflow prior to 1.23.1, describing an insecure temporary file issue. The root cause is use of the deprecated tempfile.mktemp() pattern in the affected code, with remediation to upgrade to mlflow 1.23.1 or later as indicated by OSV/GHSA entries. The connected sources c...

CVE-2023-6909

Mlflow up to version 2.9.2 is affected by CVE-2023-6909: a path traversal in the repository mlflow/mlflow allows escaping to read sensitive files via the sequence \..\filename. The vulnerability affects the mlflow/mlflow project prior to 2.9.2 and is classified as CWE-29. Impact in the NVD/NVD-de...

CVE-2026-4035

CVE-2026-4035 affects mlflow/mlflow versions before 3.11.0. The API for AI Gateway secrets allows the api_key field to contain $ENV_VAR references, which are resolved against the MLflow server environment at runtime. Attackers can exfiltrate server-side environment credentials (e.g., AWS_ACCESS_K...

CVE-2023-6018

CVE-2023-6018 affects MLflow. An unauthenticated attacker can overwrite any file on the server hosting MLflow, potentially compromising integrity and enabling remote code execution per connected documents. Reported impact emphasizes unauthenticated file writes with high severity (CVSS up to 10.0 ...

CVE-2025-0453

CVE-2025-0453 : In mlflow/mlflow version 2.17.2, the /graphql endpoint is vulnerable to denial of service via large batches of queries that repeatedly request all runs from a given experiment, causing uncontrolled resource consumption and making the application unresponsive. The issue is document...

CVE-2024-8859

Mlflow/mlflow 2.15.1 contains a path traversal/local file read vulnerability when using the dbfs service: the URL is interpolated into the file protocol with only the path portion validated, enabling reading arbitrary server files when dbfs is mounted locally. Public sources (Nuclei template, OSV...

CVE-2024-4263

CVE-2024-4263 describes a broken access control in mlflow/mlflow prior to 2.10.1, where users with EDIT permissions on an experiment can delete artifacts they should only be able to read/update. The issue stems from insufficient validation of DELETE requests for artifact deletions, enabling unaut...

CVE-2024-27132

MLflow suffers from insufficient sanitization of template variables, enabling XSS when loading an untrusted recipe and potentially enabling client-side RCE in Jupyter Notebook. The root cause is lack of input sanitization in rendering templates. Public details about affected versions or patches a...

CVE-2023-1176

The CVE affects the open source MLflow project mlflow/mlflow prior to version 2.2.2 (ABSOLUTE PATH TRAVERSAL). Documented in multiple sources, the vulnerability allows an attacker to traverse the filesystem to access arbitrary files on the host via the mlflow server/ui workflow (attack vector: LO...

CVE-2023-2356

CVE-2023-2356 is a Relative Path Traversal flaw in mlflow/mlflow before 2.3.1. The connected sources (Nuclei template, OSV entries, GHSA advisories) describe a vulnerability allowing traversal to read sensitive server files (LFI) via a crafted path. Affected software: mlflow/mlflow, versions prio...

CVE-2023-3765

MLflow (mlflow/mlflow) prior to version 2.5.0 contains an Absolute Path Traversal vulnerability. The issue arises in an MLflow repository and can lead to unauthorized access to sensitive information stored on the server. According to the connected sources, the affected component is mlflow/mlflow’...

CVE-2023-6015

CVE-2023-6015 corresponds to MLflow where the server could be subjected to an arbitrary file upload. Multiple connected sources corroborate: the vulnerability is described as allowing arbitrary files to be PUT onto the MLflow server, with various third-party advisories and vulnerability databases...

CVE-2024-0520

CVE-2024-0520 affects mlflow/mlflow v8.2.1, enabling remote code execution via command injection in mlflow.data.http_dataset_source.py when loading an HTTP URL dataset. The filename gathered from Content-Disposition or URL path is used to form the final file path without proper sanitization, allo...

CVE-2023-6568

MLflow XSS (CVE-2023-6568) : A reflected XSS exists in mlflow/mlflow due to how the Content-Type header from POST requests is handled. The vulnerability is in mlflow/server/auth/init .py, where user-supplied Content-Type is directly inserted into a Python-formatted string and returned, allowing a...

CVE-2023-6977

CVE-2023-6977 relates to Mlflow prior to 2.8.0 suffering a local file inclusion (path traversal) vulnerability. The Nuclei template specifies that an attacker could read sensitive files on the server and potentially modify data or perform unauthorized admin operations in context of the affected s...

CVE-2024-27134

CVE-2024-27134 : Multiple connected sources confirm a vulnerability in MLflow’s spark_udf API where excessive directory permissions allow a local attacker to achieve privilege escalation via a ToCToU attack. Affected: MLflow (spark_udf path) with local execution context. Root cause: insufficient ...

CVE-2024-37052

The CVE-2024-37052 issue affects MLflow platforms running version 1.1.0 or newer. The root cause is deserialization of untrusted data during model handling, as described by multiple connected sources, including reports that an uploaded scikit-learn model can be crafted to trigger arbitrary code e...

CVE-2023-6831

mlflow/mlflow prior to 2.9.2 is exposed to a path traversal in artifact deletion. The root cause is a double decoding path handling: _delete_artifact_mlflow_artifacts and local_file_uri_to_path allow traversal due to an extra unquote operation in delete_artifacts. This can enable deletion of arbi...

CVE-2024-1483

Summary: CVE-2024-1483 is a path traversal vulnerability in mlflow/mlflow 2.9.2 that allows an attacker to access arbitrary server files. The issue stems from insufficient validation of user-supplied input in server handlers, enabling traversal via crafted HTTP POST requests using crafted artifac...

CVE-2025-1474

Summary: CVE-2025-1474 affects mlflow/mlflow 2.18, where an admin can create a new user account without a password, potentially enabling unauthorized access. The issue is fixed in version 2.19.0. Reports across multiple sources (Red Hat, CIRCL, GHSA, osv, NVD, OSV) corroborate the same descriptio...

CVE-2025-15379

Summary: CVE-2025-15379 affects MLflow (model serving container initialization). In the function _install_model_dependencies_to_env(), when deploying with env_manager=LOCAL, dependency specs from the model artifact's python_env.yaml are interpolated into a shell command without sanitization, enab...

CVE-2024-27133

CVE-2024-27133 : Affects MLflow. Insufficient sanitization of dataset table fields in MLflow recipes can cause a client-side XSS, which in turn can lead to a client-side RCE when running the recipe in Jupyter Notebook . Root cause: lack of input sanitization for untrusted datasets in the data tab...

CVE-2024-1560

CVE-2024-1560 affects mlflow/mlflow prior to 2.9.2, in the artifact deletion path. A double decoding flaw in _delete_artifact_mlflow_artifacts and local_file_uri_to_path, via an extra unquote in delete_artifacts, allows path traversal and deletion of arbitrary server directories. Impact: high, wi...

CVE-2024-2928

Summary: MLflow

CVE-2023-2780

CVE-2023-2780 affects the open-source project mlflow/mlflow prior to version 2.3.1. A path traversal flaw, described as a Local File Inclusion bypass via the payload "..\filename", can allow an attacker to read local files on the server. The Nuclei/NVD entries consistently reference this vulnerab...

CVE-2024-1558

CVE-2024-1558 (mlflow/mlflow) describes a path traversal vulnerability in the function _create_model_version() in server/handlers.py due to improper validation of the source parameter. Attackers can bypass the check in _validate_non_local_source_contains_relative_paths(source) and gain arbitrary ...

CVE-2024-1594

CVE-2024-1594 is a path traversal vulnerability in mlflow/mlflow related to handling of artifact_location when creating an experiment. The connected OSV entry states that a fragment component # in the artifact URL can be used to bypass validation and allow reading arbitrary files on the server wi...

CVE-2026-0596

The CWE/CVE describes a command-injection in mlflow/mlflow when serving a model with enable_mlserver=True. The vulnerability occurs because model_uri is embedded directly into a shell command executed via bash -c without sanitization, allowing shell metacharacters (e.g., $(), backticks) to enable...

CVE-2024-37054

CVE-2024-37054 affects MLflow and stems from a deserialization issue in mlflow.pyfunc.load_model that processes PyFunc models, allowing a maliciously crafted model to trigger arbitrary code execution when loaded or interacted with. Public details (NVD entry) indicate versions 0.9.0 and newer are ...

CVE-2023-30172

CVE-2023-30172 describes a directory traversal in the mlflow platform’s /get-artifact API, allowing an attacker to read arbitrary server files via the path parameter. Affected: mlflow up to v2.0.1. Underlying cause: directory traversal in the get-artifact endpoint. Impact is high on confidentiali...

CVE-2024-3573

The CVE-2024-3573 entry concerns mlflow/mlflow with a Local File Inclusion (LFI) caused by improper parsing of URIs in the is_local_uri logic. The issue misclassifies URIs with empty or file schemes as non-local, enabling an attacker to craft malicious model versions (source parameter) that bypas...

CVE-2025-11201

CVE-2025-11201 is a Directory Traversal leading to Remote Code Execution in MLflow Tracking Server. The flaw arises from improper validation of a user-supplied model file path, allowing an attacker to execute code with the service account via crafted paths over the network without authentication....

CVE-2025-1473

In MLflow (mlflow/mlflow), a CSRF vulnerability affects versions 2.17.0 to 2.20.1 in the Signup feature, allowing an attacker to create a new account and potentially perform unauthorized actions on behalf of the attacker’s account. The CVE-2025-1473 entry documents the flaw and its impact as Cros...

CVE-2026-2651

MLflow CVE-2026-2651 describes missing authorization validation for MPU endpoints under /mlflow-artifacts/mpu/* when serve-artifacts is enabled. Vulnerable in MLflow versions

CVE-2025-14287

Summary: CVE-2025-14287 is a command-injection in mlflow/mlflow prior to v3.7.0. The flaw resides in mlflow/sagemaker/init .py (lines 161–167) where user-supplied container image names are directly interpolated into shell commands and executed with os.system(), enabling arbitrary command executio...

CVE-2026-2734

Summary : For mlflow/mlflow up to version 3.9.0, the REST endpoint GET /api/2.0/mlflow/model-versions/search and the GraphQL query mlflowSearchModelVersions lack per-model authorization when basic auth is enabled. This results in any authenticated user being able to enumerate all model versions a...

CVE-2023-6014

CVE-2023-6014 affects MLflow and describes an authentication bypass that lets an attacker arbitrarily create accounts via the MLflow server/UI without credentials. Documentation in connected sources confirms the vulnerability stems from bypassing MLflow’s authentication mechanism (basic auth path...

CVE-2024-1593

This CVE describes a path traversal vulnerability in the mlflow/mlflow repository caused by improper handling of URL parameters. Attackers can smuggle path traversal sequences using the ';' character in URLs to manipulate the 'params' portion and access unauthorized files or directories. The repo...

CVE-2026-10803

MLflow up to 3.10.0 contains a flaw in mlflow.data.digest_utils (Digest Computation) where manipulation leads to use of a weak hash. This affects the Digest Utils function in the Dataset Digest Computation component and enables a local attack. The reported exploitability is high in complexity wit...

CVE-2026-3198

MLflow 3.9.0 with basic-auth fails authorization for multiple Gateway API 'list' endpoints. The BEFORE_REQUEST_HANDLERS dictionary in mlflow/server/auth/init .py lacks entries for ListGatewaySecretInfos, ListGatewayEndpoints, and ListGatewayModelDefinitions, allowing any authenticated user to enu...

CVE-2026-2611

Vulnerability summary (CVE-2026-2611) : In MLflow 3.9.0, the MLflow Assistant’s /ajax-api endpoints had improper origin validation, allowing remote attackers to bypass the loopback-only restriction via cross-origin requests from malicious pages. This could let an attacker interact with the MLflow...

CVE-2026-0545

In mlflow/mlflow, the FastAPI endpoints under /ajax-api/3.0/jobs/* are unprotected when the basic-auth app is enabled. If job execution is enabled (MLFLOW_SERVER_ENABLE_JOB_EXECUTION=true) and any job function is allowlisted, any network client can submit, read, search, and cancel jobs without cr...

CVE-2026-2652

MLflow MLflow

CVE-2023-43472

MLFlow before 2.8.1 is affected by CVE-2023-43472. A remote attacker can disclose sensitive information via a crafted request to the MLFlow REST API. Impact described in sources: access to sensitive information stored in MLFlow. Root cause: issue exists in MLFlow 2.8.1 and earlier as stated in th...

CVE-2025-15036

CVE-2025-15036 concerns a path-traversal vulnerability in the mlflow/mlflow project, specifically in mlflow/pyfunc/dbconnect_artifact_cache.py within extract_archive_to_dir(). The issue arises from unvalidated tar member paths during extraction, allowing a tar.gz controlled by an attacker to over...