Kubeflow is vulnerable to the theft of authorization tokens by any user of the
Kubeflow UI or APIs, such as the Dashboard, Pipelines API, or Notebooks. With
this token, the attacker can take over the user’s account and the data that is
processed by that user. The attacker needs a valid user with the kubeflow-edit
or Contributor role in a random Kubeflow namespace to perform this attack. This
is given if Automatic Profile Creation is enabled. A setup based on the
official manifests prior to version 1.10, and on most other packaged Kubeflow
distributions, is vulnerable.
The Istio edit permissions were removed by Kubeflow in a timely manner. Affected
users should update to the latest version to mitigate this issue.
Hardening a Linux client system to an acceptable degree is a time-consuming
process, one that demands familiarity with a broad set of configuration
parameters, framework recommendations, and the reasoning behind each control.
This post introduces our new Linux client hardening guide
(MD,
PDF), a comprehensive, publicly
available hardening reference for Linux systems.
Motivation and Scope
The guide covers the full breadth of controls needed to significantly raise the
security posture of a modern Linux installation while preserving operational
usability (this will be very subjective, the guide reflects my opinion of
“usable”). It has been developed and validated against Ubuntu 24.04 LTS as the
primary reference platform, and cross-tested on Fedora, Debian 12, and Arch
Linux as well as on traditionally server-oriented distributions like openSUSE
Leap 15.6, Debian 12, Rocky Linux 9, and Red Hat Enterprise Linux 9 while not
focussing on those as the guide is created for Linux clients.
Over the last few weeks, I have had a very productive exchange with
Christoph Klaassen
on the impact of AI on security governance and compliance. In this post, we
summarize our thoughts.
While investigating how process mitigation settings are initialized, I
encountered the global variable PspSystemMitigationOptions. Tracing how this
value is populated led me to the CmControlVector. In this blog post, we take a
look at the Windows kernel land configuration manager, especially its global
CmControlVector variable. Quick note: the kernel’s configuration manager is
not related to Microsoft Intune’s
Configuration Manager.
In short, the configuration manager is responsible for managing and implementing
the registry. However, it is also responsible for setting up parts of the system
during early boot.
Exactly one week ago, Sven and I had the incredible opportunity to give our very
first talk at KubeCon + CloudNativeCon
2026: How To Break Multi-Tenancy Again and Again …and What We Can Learn From It.
We discussed the challenges of namespace-based multi-tenancy and presented
real-world exploits in
Kubeflow, Istio,
and Traefik that bypass threat boundaries between namespaces and workloads.
Based on these problems, we developed a methodology to assess and address them.
You can find the methodology discussed in the talk in
detail in another blog post or
on GitHub. You can also find the
slides here.
This page introduces our structured methodology for assessing security risks in
Kubernetes environments that use Namespace-based Multi-Tenancy. It addresses
weaknesses that break Namespace-based isolation that not well studied, yet. We
found this issues during our research and presented them together with this
methodology in our
Talk at KubeCon + CloudNativeCon Europe 2026.
The methodology assumes that industry best practices, such as NetworkPolicies,
Role-Based Access Control (RBAC), and Pod Security Standards, are already in
place. These measures provide a necessary baseline level of protection against
well-known isolation threats. However, they are insufficient to address a class
of more subtle attack vectors arising from interactions between tenants and
shared components. Such attack vectors may still compromise the confidentiality,
integrity, and availability (CIA) of the cluster and its workloads, even in
well-hardened environments.
We reported a possible Man-in-the-Middle (MitM) attack scenario in which
a VirtualService can redirect or intercept traffic within the service mesh.
This affects Namespace-based Multi-Tenancy clusters where tenants have the
permissions to deploy Istio resources (networking.istio.io/v1).
This blog post highlights the risks of using Istio in multi-tenant clusters and
explains how users can mitigate these risks and safely operate Istio in their
deployments.
There is a growing landscape of security products promising to protect an
organization’s IT infrastructure from attacks. Solutions referred to as EDR, and
sometimes also as XDR, are designed to protect endpoints from all malicious
activity. The ever-increasing cases of breaches and the associated costs,
especially in the realm of
ransomware attacks,
raise the question of whether there is more that can be done to add an
additional layer to traditional endpoint protection concepts. That is why a
customer of ours commissioned us to evaluate whether EDR supplementing solutions
provide extended protection against ever-evolving threats, as well as to shine a
light on the performance overheads those solutions might introduce.
When conducting pentests of Bluetooth devices or whilst working on Bluetooth
related research, we often use Bumble. In
this Blogpost I will present a solution to capture a live stream of Bumble
Bluetooth traffic in Wireshark.
Bumble is a fully featured Bluetooth stack, written entirely in Python. What
makes it extremely powerful for security assessments and research is the level
of control it provides. It can simulate certain conditions, including errors,
with a level of precision that most Bluetooth stacks don’t offer. However,
sometimes you not only need control, you also need visibility.