/

CVE-2026-50566

Severity:

CRITICAL RISK

(9.9)

Privilege Misconfiguration (CWE-250) in Fission RBAC

CVE-2026-50566

Deferred

Vulnerability type

Category:

Privilege Management Misconfiguration

CWE:

CWE-250, CWE-269

Impact:

Container Escape and Cluster Compromise

Severity:

CRITICAL RISK

(9.9)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A privilege misconfiguration in Fission allows tenants with specific RBAC permissions to schedule highly privileged containers using the executor's powerful service account.

When exploited, this can lead to container sandbox escapes, host filesystem and network access, and possible node- or cluster-level compromise in affected Kubernetes environments.

Affected context

This vulnerability occurs in Kubernetes clusters running Fission versions prior to 1.24.0 where:

  • Tenants are granted environments.fission.io create/update permissions via RBAC
  • The Fission executor operates with a high-privilege Kubernetes service account
  • Function or builder pods run in namespaces controlled or influenced by tenants

Under these conditions, a tenant can define pod specs that request elevated privileges, which the executor then runs using its own powerful identity, blurring the boundary between tenant and cluster-level permissions.

Exploit preconditions

  • Attack surface: Kubernetes API via Fission's environment and function configuration resources
  • Required access:
  • Ability to authenticate to the cluster
  • RBAC permissions to create or update environments.fission.io resources
  • Trigger point: Submission or update of an environment or function configuration that specifies privileged, allowPrivilegeEscalation, or dangerous Linux capabilities
  • Why exploitation is possible: The Fission executor honors tenant-specified container security settings but runs workloads under a high-privilege service account, effectively elevating tenant-controlled pods beyond their intended isolation.

Why this happens

The root cause lies in how Fission couples tenant-influenced pod specifications with a high-privilege executor service account without sufficiently constraining security-sensitive fields.

Tenants with environments.fission.io create/update rights can declare privileged containers, enable allowPrivilegeEscalation, or request powerful Linux capabilities, which are then started by the executor using its elevated Kubernetes identity.

This aligns with:

  • CWE-250: Execution with Unnecessary Privileges, because workloads run with more privileges than required.
  • CWE-269: Improper Privilege Management, since privilege boundaries between tenants and the cluster service account are not strictly enforced.

The combination of over-privileged service accounts and insufficient policy controls on user-specified pod security options creates a path for privilege escalation.

Potential impact

  • Container sandbox escape from Fission function or builder pods
  • Read and write access to the host filesystem from compromised pods
  • Direct access to host networking and potentially other cluster workloads
  • Lateral movement across nodes via compromised Kubernetes credentials or mounts
  • Complete node compromise and potential full cluster takeover
  • Abuse of cluster resources for data exfiltration, cryptomining, or further attacks

General mitigation guidance

Effective mitigation focuses on tightening privilege boundaries and restricting what tenants can specify in pod configurations:

  • Reduce executor service account privileges to the minimum necessary for scheduling and managing function workloads.
  • Enforce Pod Security Standards or equivalent admission policies to block privileged pods, allowPrivilegeEscalation, and dangerous capabilities in Fission-managed namespaces.
  • Limit RBAC grants so that tenants cannot arbitrarily create or update environments.fission.io with unconstrained security context fields.
  • Separate trust levels into different namespaces and service accounts, preventing low-trust tenants from sharing high-privilege runtimes.
  • Apply default-deny policies for privileged operations and require explicit, tightly controlled exceptions for operational needs.

How teams detect issues like this

Detection can combine configuration review, policy enforcement, and runtime monitoring:

  • Use static and policy-as-code checks (e.g., OPA/Gatekeeper, Kyverno) to flag Fission environments or functions that request privileged: true, allowPrivilegeEscalation: true, or elevated capabilities.
  • Periodically audit RBAC roles and role bindings that grant environments.fission.io create/update permissions, focusing on multi-tenant contexts.
  • Review service accounts used by the Fission executor and associated ClusterRole/Role definitions for excessive permissions.
  • Monitor Kubernetes audit logs for the creation of privileged pods in Fission namespaces or pods running under sensitive service accounts.
  • Employ container security platforms or runtime detection tools to alert on host mount usage, privilege escalation attempts, and anomalous access originating from Fission-managed pods.

Sources

SourceLink
github.comgithub.com/fission/fission/pull/3406
github.comgithub.com/fission/fission/releases/tag/v1.24.0
github.comgithub.com/fission/fission/security/advisories/GHSA-m63v-2g9w-2w6v

Frequently asked

questions

It is exploitable by authenticated tenants who can reach the Kubernetes API or Fission control plane and have environments.fission.io create/update permissions. It is not typically exploitable by unauthenticated external users unless your cluster API or Fission endpoints are exposed without proper access control.

The attacker must be a Kubernetes tenant or user with RBAC permissions to create or modify Fission environment definitions, and the ability to submit function or builder configurations that Fission will schedule.

The safest pattern is to decouple tenant-controlled configurations from highly privileged service accounts by reducing executor privileges, enforcing strict pod security policies, and constraining which security-related fields tenants can set.

Design multi-tenant systems around least-privilege principles, apply admission controls for pod security, regularly audit RBAC and service account permissions, and treat any user-influenced pod spec fields as untrusted input that must pass through strong policy and validation layers.

Find out what's already
exploitable in your codebase.