DEBIAN-CVE-2025-9905

The Keras Model.loadmodel method can be exploited to achieve arbitrary code execution, even with safemode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.loadmodel, will trigger arbitrary code to be executed. This is achieved by crafting a special...

Improper Control of Dynamically-Managed Code Resources

Overview keras is a Keras is a high-level neural networks API for Python.. Affected versions of this package are vulnerable to Improper Control of Dynamically-Managed Code Resources via the Model.loadmodel method. An attacker can execute arbitrary code by providing a specially crafted .h5 or .hdf...

CVE-2025-9905 Arbitary Code execution in Keras load_model()

The Keras Model.loadmodel method can be exploited to achieve arbitrary code execution, even with safemode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.loadmodel, will trigger arbitrary code to be executed. This is achieved by crafting a special...

CVE-2025-9905 Arbitary Code execution in Keras load_model()

The Keras Model.loadmodel method can be exploited to achieve arbitrary code execution, even with safemode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.loadmodel, will trigger arbitrary code to be executed. This is achieved by crafting a special...

CVE-2025-9905

CVE-2025-9905 affects Keras Model.load_model for .h5/.hdf5 archives, where safe_mode=True is ignored and Lambda layer can trigger arbitrary code execution. Vulnerability details: loading a crafted .h5 can execute code via pickled Lambda; legacy .h5 format is involved. Remediation: upgrade Keras t...

CVE-2025-9905 Arbitary Code execution in Keras load_model()

The Keras Model.loadmodel method can be exploited to achieve arbitrary code execution, even with safemode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.loadmodel, will trigger arbitrary code to be executed. This is achieved by crafting a special...

CVE-2025-9906

CVE-2025-9906 affects Keras Model.load_model: loading a specially crafted .keras archive can bypass safe_mode and trigger arbitrary code execution via a config.json entry that enables unsafe deserialization, followed by a Lambda layer with pickled code. Impact is arbitrary code execution during m...

CVE-2025-9906 Arbitrary Code execution in Keras Safe Mode

The Keras Model.loadmodel method can be exploited to achieve arbitrary code execution, even with safemode=True. One can create a specially crafted .keras model archive that, when loaded via Model.loadmodel, will trigger arbitrary code to be executed. This is achieved by crafting a special...

CVE-2025-9906 Arbitrary Code execution in Keras Safe Mode

The Keras Model.loadmodel method can be exploited to achieve arbitrary code execution, even with safemode=True. One can create a specially crafted .keras model archive that, when loaded via Model.loadmodel, will trigger arbitrary code to be executed. This is achieved by crafting a special...

PT-2025-38517

Name of the Vulnerable Software and Affected Versions Keras affected versions not specified Description The Model.load model method can be exploited to achieve arbitrary code execution, even when safe mode is enabled. This is possible by creating a specially crafted .h5 or .hdf5 model archive tha...

Keras 安全漏洞

Keras is a multi-backend deep learning framework open-sourced by Keras. A security vulnerability exists in Keras that stems from not properly enforcing the safemode option, which could lead to the execution of arbitrary code...

How CISOs Can Drive Effective AI Governance

AI’s growing role in enterprise environments has heightened the urgency for Chief Information Security Officers CISOs to drive effective AI governance. When it comes to any emerging technology, governance is hard – but effective governance is even harder. The first instinct for most organizations...

How CISOs Can Drive Effective AI Governance

AI's growing role in enterprise environments has heightened the urgency for Chief Information Security Officers CISOs to drive effective AI governance. When it comes to any emerging technology, governance is hard – but effective governance is even harder. The first instinct for most organizations...

CVE-2025-59518

In LemonLDAP::NG before 2.16.7 and 2.17 through 2.21 before 2.21.3, OS command injection can occur in the Safe jail. It does not Localize during rule evaluation. Thus, an administrator who can edit a rule evaluated by the Safe jail can execute commands on the server...

DEBIAN-CVE-2025-59518

In LemonLDAP::NG before 2.16.7 and 2.17 through 2.21 before 2.21.3, OS command injection can occur in the Safe jail. It does not Localize during rule evaluation. Thus, an administrator who can edit a rule evaluated by the Safe jail can execute commands on the server...

UBUNTU-CVE-2025-59518

In LemonLDAP::NG before 2.16.7 and 2.17 through 2.21 before 2.21.3, OS command injection can occur in the Safe jail. It does not Localize during rule evaluation. Thus, an administrator who can edit a rule evaluated by the Safe jail can execute commands on the server...

CVE-2025-59518

In LemonLDAP::NG before 2.16.7 and 2.17 through 2.21 before 2.21.3, OS command injection can occur in the Safe jail. It does not Localize during rule evaluation. Thus, an administrator who can edit a rule evaluated by the Safe jail can execute commands on the server...

CVE-2025-59518

In LemonLDAP::NG before 2.16.7 and 2.17 through 2.21 before 2.21.3, OS command injection can occur in the Safe jail. It does not Localize during rule evaluation. Thus, an administrator who can edit a rule evaluated by the Safe jail can execute commands on the server...

PT-2025-38108

Name of the Vulnerable Software and Affected Versions: LemonLDAP::NG versions prior to 2.16.7 LemonLDAP::NG versions 2.17 through 2.21 before 2.21.3 Description: LemonLDAP::NG is susceptible to OS command injection within the Safe jail. The software fails to localize the underscore character duri...

CVE-2025-59518

CVE-2025-59518 affects LemonLDAP::NG versions prior to 2.16.7 and 2.17 through 2.21 before 2.21.3, where the Safe jail can be exploited for OS command injection. The root cause is that Safe jail rules are not localized during evaluation, allowing an administrator who edits a rule to execute comma...