Live CVE & 0day intelligence

Sound the depths of every vulnerability.

Newly released CVEs across every platform — sleek to read, verbose on demand, and served raw as markdown for AI and agent ingestion. Severity reads as depth: the deeper the contact, the graver the threat.

Browse the feed For agents → /llms.txt

Depth = severity + exploitation

HadalCritical and actively exploited (CISA KEV / 0day)
AbyssalCritical with a public exploit or in-the-wild use
MidnightCritical, or high with PoC / in-the-wild
TwilightHigh severity, or a signal on a lesser flaw
SunlitLow / medium · no exploitation signal

31673

Tracked CVEs

3155

Critical

385

Exploited in wild

3811

Last 7 days

Trending now

Bluesky + Hacker News mentions, last 7 days

1CVE-2026-85046High· 8.8Type confusion in V8 in Google Chrome prior to 152.0.7977.82 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page874
2CVE-2026-58704High· 8.8In Cellular Modem, there is a possible permission bypass due to a logic error in the code373
3CVE-2026-69836Critical· 10.0Microsoft Entra ID Remote Code Execution Vulnerability267
4CVE-2026-66747Critical· 9.8Zbtlink router firmware ships an embedded remote-control implant, ENDLESSDOORS, present in every published build across the product line266
5CVE-2026-18963Critical· 9.1A flaw was found in the reset-credentials flow of the keycloak-services component, which is the core engine for identity and access management in Red Hat Build of Keycloak263

This week's digest →All digests

Feed

51 matches

CVE-2026-96804High· 8.8

yesterday

MLflow's statsmodel flavor, versions 2.1.0 to 3.14.0, omits the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control entirely in _load_model(), which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

MLflow's statsmodel flavor, versions 2.1.0 to 3.14.0, omits the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control entirely in _load_model(), which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

▾ TwilightMLflow · MLflowvia NVD

CVE-2026-96775High· 8.8

yesterday

MLflow's dspy flavor, versions >= 2.0, applies the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control only when the model_path ends in .pkl, which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

MLflow's dspy flavor, versions >= 2.0, applies the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control only when the model_path ends in .pkl, which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

▾ TwilightMLflow · MLflowvia NVD

CVE-2026-79721High· 8.6

2w ago

Code execution can occur in versions of the MLflow platform running version 0.0.1 or newer, enabling a maliciously crafted model artifact to execute arbitrary code on an end user's system when loaded by the project.

Code execution can occur in versions of the MLflow platform running version 0.0.1 or newer, enabling a maliciously crafted model artifact to execute arbitrary code on an end user's system when loaded by the project.

▾ Twilightmlflow · mlflowEPSS 0.29%via NVD

GHSA-gqvg-gmmx-x4hmHigh· 8.8

3w ago

MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor — RCE via crafted model artifact

▾ Twilightmlflow · mlflowvia GHSA

CVE-2026-69146Medium· 6.5

1mo ago

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. From 3.13.0 until 3.15.0, LogInputs is absent from BEFORE_REQUEST_HANDLERS in the mlflow/server/auth package, allowing any a…

▾ Sunlitmlflow · mlflowEPSS 0.28%via NVD

CVE-2026-69148High· 7.1

1mo ago

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. Prior to 3.15.0, CreateModelVersion accepts a run_id or model_id after _validate_source_run() or _validate_source_model() in…

▾ Twilightmlflow · mlflowEPSS 0.28%via NVD

CVE-2026-71211High· 7.1PoC

1mo ago

MLflow's AI Gateway accepts an auth_config.api_base value when creating a gateway secret (mlflow/server/handlers.py, _create_gateway_secret) with no validation of scheme, host, or IP range; the value is stored verbatim

MLflow's AI Gateway accepts an auth_config.api_base value when creating a gateway secret (mlflow/server/handlers.py, _create_gateway_secret) with no validation of scheme, host, or IP range; the value is stored verbatim. The gateway proxy…

▾ Midnightmlflow · mlflowEPSS 0.25%via NVD

CVE-2026-8147High· 8.1

2mo ago

MLflow: trace API endpoints lack proper authorization validators

▾ Twilightmlflow · mlflowEPSS 0.55%via OSV

CVE-2026-10803Low· 3.6

3mo ago

MLflow: Deterministic sampling in dataset digest enables predictable collisions

▾ Sunlitmlflow · mlflowEPSS 0.10%via OSV

CVE-2026-4035Critical· 9.1

3mo ago

MLflow: Environment variable injection in AI Gateway secrets enables server-side credential exfiltration

▾ Midnightmlflow · mlflowEPSS 0.52%via OSV

CVE-2026-3198Medium· 6.5

3mo ago

MLflow: Any authenticated user can enumerate all gateway secrets, endpoints, and model definitions

▾ Sunlitmlflow · mlflowEPSS 0.24%via OSV

CVE-2026-2651Critical· 9.0

4mo ago

MLflow allows unauthorized access to multipart upload endpoints when the `--serve-artifacts` mode is enabled

▾ Midnightmlflow · mlflowEPSS 0.34%via OSV

CVE-2026-2734Medium· 6.5

4mo ago

MLflow authenticated users can enumerate any registered model versions due to lack of per-model permissions checks

▾ Sunlitmlflow · mlflowEPSS 0.44%via OSV

CVE-2026-4137High· 7.0

4mo ago

MLflow has a command injection in mlflow/sagemaker/init.py

MLflow has a command injection in mlflow/sagemaker/__init__.py

▾ Twilightmlflow · mlflowEPSS 1.5%via OSV

CVE-2026-2033High· 8.10day

7mo ago

MLflow Tracking Server Artifact Handler Directory Traversal Remote Code Execution Vulnerability

▾ Abyssalmlflow · mlflowEPSS 1.7%via OSV

CVE-2026-2635High· 7.30day⚖ disputed

7mo ago

MLflow Use of Default Password Authentication Bypass Vulnerability

MLflow Use of Default Password Authentication Bypass Vulnerability. This vulnerability allows remote attackers to bypass authentication on affected installations of MLflow. Authentication is not required to exploit this vulnerability. T…

▾ AbyssalMLflow · MLflowEPSS 0.98%via NVD

CVE-2025-10279High· 7.0

7mo ago

mlflow Creates of Temporary File in Directory with Insecure Permissions

▾ Twilightmlflow · mlflowEPSS 0.23%via OSV

CVE-2025-14279High· 8.1

8mo ago

MLFlow is vulnerable to DNS rebinding attacks due to a lack of Origin header validation

▾ Twilightmlflow · mlflowEPSS 0.21%via OSV

From the blog

All posts →

Sep 8, 2026

Sound the depths of every vulnerability.

Trending now

Feed

MLflow's statsmodel flavor, versions 2.1.0 to 3.14.0, omits the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control entirely in _load_model(), which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

MLflow's dspy flavor, versions >= 2.0, applies the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control only when the model_path ends in .pkl, which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

Code execution can occur in versions of the MLflow platform running version 0.0.1 or newer, enabling a maliciously crafted model artifact to execute arbitrary code on an end user's system when loaded by the project.

MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor — RCE via crafted model artifact

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models

MLflow's AI Gateway accepts an auth_config.api_base value when creating a gateway secret (mlflow/server/handlers.py, _create_gateway_secret) with no validation of scheme, host, or IP range; the value is stored verbatim

MLflow: trace API endpoints lack proper authorization validators

MLflow: Deterministic sampling in dataset digest enables predictable collisions

MLflow: Environment variable injection in AI Gateway secrets enables server-side credential exfiltration

MLflow: Any authenticated user can enumerate all gateway secrets, endpoints, and model definitions

MLflow allows unauthorized access to multipart upload endpoints when the `--serve-artifacts` mode is enabled

MLflow authenticated users can enumerate any registered model versions due to lack of per-model permissions checks

MLFlow Creates a Temporary File With Insecure Permissions

MLflow: unauthenticated access to certain FastAPI routes

MLflow Has a Server-Side Request Forgery (SSRF) Vulnerability

MLflow is vulnerable to an authorization bypass affecting the AJAX endpoint

MLFlow path traversal vulnerability

MLFlow allows Tracing + Assessments Access

Arbitrary file write via tar traversal in mlflow

MLflow has a command injection in mlflow/sagemaker/__init__.py

MLflow Tracking Server Artifact Handler Directory Traversal Remote Code Execution Vulnerability

MLflow Use of Default Password Authentication Bypass Vulnerability

mlflow Creates of Temporary File in Directory with Insecure Permissions

MLFlow is vulnerable to DNS rebinding attacks due to a lack of Origin header validation

From the blog

What's new: open corpus, three new MCP tools, vendor advisories, package pages

Wiring VulnSea into an agent: a hands-on tutorial

VulnSea is live

Sound the depths of every vulnerability.

Trending now

Feed

MLflow's statsmodel flavor, versions 2.1.0 to 3.14.0, omits the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control entirely in _load_model(), which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

MLflow's dspy flavor, versions >= 2.0, applies the MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False security control only when the model_path ends in .pkl, which allows a remote attacker to execute arbitrary code via a crafted MLmodel artifact.

Code execution can occur in versions of the MLflow platform running version 0.0.1 or newer, enabling a maliciously crafted model artifact to execute arbitrary code on an end user's system when loaded by the project.

MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor — RCE via crafted model artifact

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models

MLflow's AI Gateway accepts an auth_config.api_base value when creating a gateway secret (mlflow/server/handlers.py, _create_gateway_secret) with no validation of scheme, host, or IP range; the value is stored verbatim

MLflow: trace API endpoints lack proper authorization validators

MLflow: Deterministic sampling in dataset digest enables predictable collisions

MLflow: Environment variable injection in AI Gateway secrets enables server-side credential exfiltration

MLflow: Any authenticated user can enumerate all gateway secrets, endpoints, and model definitions

MLflow allows unauthorized access to multipart upload endpoints when the `--serve-artifacts` mode is enabled

MLflow authenticated users can enumerate any registered model versions due to lack of per-model permissions checks

MLFlow Creates a Temporary File With Insecure Permissions

MLflow: unauthenticated access to certain FastAPI routes

MLflow Has a Server-Side Request Forgery (SSRF) Vulnerability

MLflow is vulnerable to an authorization bypass affecting the AJAX endpoint

MLFlow path traversal vulnerability

MLFlow allows Tracing + Assessments Access

Arbitrary file write via tar traversal in mlflow

MLflow has a command injection in mlflow/sagemaker/__init__.py

MLflow Tracking Server Artifact Handler Directory Traversal Remote Code Execution Vulnerability

MLflow Use of Default Password Authentication Bypass Vulnerability

mlflow Creates of Temporary File in Directory with Insecure Permissions

MLFlow is vulnerable to DNS rebinding attacks due to a lack of Origin header validation

From the blog

What's new: open corpus, three new MCP tools, vendor advisories, package pages

Wiring VulnSea into an agent: a hands-on tutorial

VulnSea is live

MLflow has a command injection in mlflow/sagemaker/init.py

MLflow has a command injection in mlflow/sagemaker/init.py