GHSA-756X-M4MJ-Q96C
Vulnerability from github – Published: 2025-01-30 17:52 – Updated: 2025-02-05 16:24Impact
The policy group feature, added to by the 1.17.0 release, introduced two new types of CRD: ClusterAdmissionPolicyGroup and AdmissionPolicyGroup. The former is cluster wide, while the latter is namespaced.
By being namespaced, the AdmissionPolicyGroup has a well constrained impact on cluster resources. Hence, it’s considered safe to allow non-admin users to create and manage these resources in the namespaces they own. Kubewarden policies can be allowed to query the Kubernetes API at evaluation time; these types of policies are called “context aware“. Context aware policies can perform list and get operations against a Kubernetes cluster. The queries are done using the ServiceAccount of the Policy Server instance that hosts the policy. That means that access to the cluster is determined by the RBAC rules that apply to that ServiceAccount. The AdmissionPolicyGroup CRD allowed the deployment of context aware policies. This could allow an attacker to obtain information about resources that are out of their reach, by leveraging a higher access to the cluster granted to the ServiceAccount token used to run the policy.
The impact of this vulnerability depends on the privileges that have been granted to the ServiceAccount used to run the Policy Server and assumes that users are using the recommended best practices of keeping the Policy Server's ServiceAccount least privileged. By default, the Kubewarden helm chart grants access to the following resources (cluster wide) only: Namespace, Pod, Deployment and Ingress.
Patches
Starting from the 1.21.0 release, the AdmissionPolicyGroup CRD does not allow the definition of context aware policies. No modifications are needed neither for performing the upgrade nor afterwards.
Workarounds
On clusters running Kubewarden < 1.21.0, the following Kubewarden policy can be applied to prevent the creation of AdmissionPolicyGroup resources that have access to Kubernetes resources:
apiVersion: policies.kubewarden.io/v1
kind: ClusterAdmissionPolicy
metadata:
name: "deny-admission-policy-groups-with-context-resources"
spec:
module: registry://ghcr.io/kubewarden/policies/cel-policy:latest
settings:
variables:
- name: hasContextAwareResources
expression: "object.spec.policies.exists(p, has(object.spec.policies[p].contextAwareResources))"
- name: isPendingDeletion
expression: "has(object.metadata.deletionTimestamp)"
validations:
- expression: "!variables.hasContextAwareResources || variables.isPendingDeletion"
message: "AdmissionPolicyGroup has contextAwareResources defined"
rules:
- apiGroups: ["policies.kubewarden.io"]
apiVersions: ["v1"]
operations: ["CREATE", "UPDATE"]
resources: ["admissionpolicygroups"]
mutating: false
backgroundAudit: true
Once the policy is applied, the Kubewarden Audit Scanner can be used to identify the AdmissionPolicyGroup policies that are violating this policy.
For more information
If you have any questions or comments about this advisory you can contact the Kubewarden team using the procedures described under the “security disclosure“ guidelines of the Kubewarden project.
{
"affected": [
{
"package": {
"ecosystem": "Go",
"name": "github.com/kubewarden/kubewarden-controller"
},
"ranges": [
{
"events": [
{
"introduced": "1.17.0"
},
{
"fixed": "1.21.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2025-24784"
],
"database_specific": {
"cwe_ids": [
"CWE-200",
"CWE-285"
],
"github_reviewed": true,
"github_reviewed_at": "2025-01-30T17:52:25Z",
"nvd_published_at": "2025-01-30T16:15:31Z",
"severity": "MODERATE"
},
"details": "### Impact\n\nThe [policy group feature](https://docs.kubewarden.io/explanations/policy-groups), added to by the 1.17.0 release, introduced two new types of CRD: ClusterAdmissionPolicyGroup and AdmissionPolicyGroup. The former is cluster wide, while the latter is namespaced.\n\nBy being namespaced, the AdmissionPolicyGroup has a well constrained impact on cluster resources. Hence, it\u2019s considered safe to allow non-admin users to create and manage these resources in the namespaces they own.\nKubewarden policies can be allowed to query the Kubernetes API at evaluation time; these types of policies are called \u201c[context aware](https://docs.kubewarden.io/reference/spec/context-aware-policies)\u201c.\nContext aware policies can perform list and get operations against a Kubernetes cluster. The queries are done using the ServiceAccount of the Policy Server instance that hosts the policy. That means that access to the cluster is determined by the RBAC rules that apply to that ServiceAccount.\nThe AdmissionPolicyGroup CRD allowed the deployment of context aware policies. This could allow an attacker to obtain information about resources that are out of their reach, by leveraging a higher access to the cluster granted to the ServiceAccount token used to run the policy.\n\n\nThe impact of this vulnerability depends on the privileges that have been granted to the ServiceAccount used to run the Policy Server and assumes that users are using the recommended best practices of keeping the Policy Server\u0027s ServiceAccount least privileged. By default, the Kubewarden helm chart grants access to the following resources (cluster wide) only: Namespace, Pod, Deployment and Ingress.\n\n### Patches\n\nStarting from the 1.21.0 release, the AdmissionPolicyGroup CRD does not allow the definition of context aware policies. No modifications are needed neither for performing the upgrade nor afterwards.\n\n### Workarounds\n\nOn clusters running Kubewarden \u003c 1.21.0, the following Kubewarden policy can be applied to prevent the creation of AdmissionPolicyGroup resources that have access to Kubernetes resources:\n\n```yaml\napiVersion: policies.kubewarden.io/v1\nkind: ClusterAdmissionPolicy\nmetadata:\n name: \"deny-admission-policy-groups-with-context-resources\"\nspec:\n module: registry://ghcr.io/kubewarden/policies/cel-policy:latest\n settings:\n variables:\n - name: hasContextAwareResources\n expression: \"object.spec.policies.exists(p, has(object.spec.policies[p].contextAwareResources))\"\n - name: isPendingDeletion\n expression: \"has(object.metadata.deletionTimestamp)\"\n validations:\n - expression: \"!variables.hasContextAwareResources || variables.isPendingDeletion\"\n message: \"AdmissionPolicyGroup has contextAwareResources defined\"\n rules:\n - apiGroups: [\"policies.kubewarden.io\"]\n apiVersions: [\"v1\"]\n operations: [\"CREATE\", \"UPDATE\"]\n resources: [\"admissionpolicygroups\"]\n mutating: false\n backgroundAudit: true\n```\n\nOnce the policy is applied, the [Kubewarden Audit Scanner](https://docs.kubewarden.io/howtos/audit-scanner) can be used to identify the AdmissionPolicyGroup policies that are violating this policy.\n\n### For more information\n\nIf you have any questions or comments about this advisory you can contact the Kubewarden team using the procedures described under the \u201c[security disclosure](https://docs.kubewarden.io/disclosure)\u201c guidelines of the Kubewarden project.",
"id": "GHSA-756x-m4mj-q96c",
"modified": "2025-02-05T16:24:28Z",
"published": "2025-01-30T17:52:25Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/kubewarden/kubewarden-controller/security/advisories/GHSA-756x-m4mj-q96c"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-24784"
},
{
"type": "WEB",
"url": "https://github.com/kubewarden/kubewarden-controller/commit/51a88dfbb4c090ce0f76a22d98106518e0824d0b"
},
{
"type": "PACKAGE",
"url": "https://github.com/kubewarden/kubewarden-controller"
},
{
"type": "WEB",
"url": "https://pkg.go.dev/vuln/GO-2025-3435"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N",
"type": "CVSS_V3"
}
],
"summary": "Kubewarden-Controller information leak via AdmissionPolicyGroup Resource"
}
Sightings
| Author | Source | Type | Date |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.