Christopher Lusk

During ABI bootstrap, the `approve-csr.service` on the bootstrap node approves and signs every Certificate Signing Request (CSR) without performing identity validation. An attacker on the installation network can retrieve the unauthenticated Ignition file via ports 22623 or 22624, extract a limited kubeconfig, and submit a CSR for `CN=system:admin`. This allows the attacker to receive a signed certificate and obtain full cluster-admin access. This issue is only exploitable during the narrow bootstrap window.

OpenShift Pipelines-as-Code (PAC) fails to verify if the creator of a Repository custom resource possesses the necessary permissions to create PipelineRuns. A user with limited permissions to create or delete Repository and Secret resources can trigger the execution of attacker-controlled PipelineRuns. These runs operate under the auto-provisioned pipeline ServiceAccount, which typically holds ClusterRole/edit permissions. This leads to namespace-scoped privilege escalation, allowing unauthorized reading of Secrets, creation, modification, or deletion of resources, and the possibility of supply chain injection within CI/CD namespaces.

An authenticated user can create projects due to a blank default configuration in the Gateway 'Create Project Role(s)' setting. This flaw allows attackers to escalate privileges to project creators and execute gateway scripts, enabling lateral movement across OT (Operational Technology) and IT networks. The issue affects critical infrastructure in the manufacturing, energy, and IT sectors worldwide.

Langflow uses a weak and reversible key derivation mechanism for encryption at rest, which could lead to the disclosure of all stored credentials.

A code injection issue exists in the `s:stepmatch()` function within the cucumber filetype plugin (`runtime/ftplugin/cucumber.vim`) for builds with +ruby support. Step-definition patterns read from `.rb` files in the `features/*/` or `stories/*/` directories are embedded into a Ruby `Kernel.eval` argument without sufficient escaping. This allows a crafted pattern in an attacker-controlled repository to execute arbitrary Ruby code and shell commands when a user invokes a step-jump mapping ([d, ]d).

A flaw exists in the Mint-mode IAM policies for AWS within the OpenShift Cloud Credential Operator. Operator credentials are provisioned with account-wide scope for destructive actions instead of being restricted to cluster-owned resources. This configuration allows for cross-scope impact if credentials are compromised.

An incomplete scope-clearing issue exists in trusted-proxy authentication mode, enabling operator.admin privilege escalation. Attackers can exploit this by declaring operator scopes on non-Control-UI clients, which allows self-declared scopes to persist on identity-bearing authentication paths.

A path traversal issue exists in the ACP dispatch component. This allows remote attackers to read arbitrary files by manipulating inbound channel attachment paths, bypassing the root directory checks and the attachment-cache.