Summary
OpenChoreo's internal cluster gateway exposed proxying, command execution, and wire-log endpoints without requiring a client certificate or authentication token. Any network-reachable caller may therefore read tenant Kubernetes Secrets, alter workloads, or execute commands across connected data planes.
Affected context
The issue occurs in OpenChoreo deployments using the internal cluster gateway before authentication is enforced on the /api/proxy/, /api/exec/, and /api/wirelogs/ endpoints. It becomes risky when the listener is reachable by an untrusted network client and the surrounding network or service configuration does not provide an independent access-control boundary.
Exploit preconditions
- Attack surface: Network-accessible internal cluster-gateway listener
- Required access: Network reachability to the exposed gateway; a valid client certificate or token is not required
- Trigger point: Requests to proxy, execution, or wire-log API endpoints
- Why exploitation is possible: The endpoints perform sensitive data-plane operations before verifying caller identity and authorization
Why this happens
The underlying flaw is a missing authentication and authorization boundary on sensitive internal API operations. Requests to endpoints capable of proxying Kubernetes actions, executing commands, or exposing logs are accepted without establishing a trusted caller identity.
This primarily maps to CWE-306: Missing Authentication for Critical Function. It also relates to CWE-862: Missing Authorization, because sensitive operations are not restricted to appropriately authorized principals, and CWE-269: Improper Privilege Management, because unauthenticated callers may inherit capabilities intended for trusted platform components.
Potential impact
- Attackers may read tenant Kubernetes Secrets
- Attackers could mutate workloads across connected data planes
- Attackers may execute commands in connected environments
- Compromised credentials or workloads could enable lateral movement
- Platform integrity, tenant isolation, and confidentiality may be reduced
General mitigation guidance
Mitigation generally focuses on enforcing identity and least privilege at every sensitive gateway operation.
Common defensive patterns include:
- Require mutually authenticated client connections or equivalent strong service identity
- Enforce token validation, audience checks, expiration, and issuer constraints where tokens are used
- Apply endpoint-specific authorization based on tenant, workload, and operation scope
- Expose internal listeners only on restricted network paths and deny untrusted lateral access
- Separate read, execution, proxy, and administrative capabilities instead of sharing broad privileges
- Run connected data-plane components with minimal Kubernetes and operating-system permissions
- Record and alert on denied requests, unusual command execution, secret access, and workload mutations
How teams detect issues like this
Organizations can identify similar weaknesses through:
- Static analysis and code review searches for internal HTTP handlers that omit authentication middleware
- Tests that invoke sensitive endpoints with no certificate, token, or identity context and expect denial
- Route and middleware inventories confirming authentication and authorization coverage for every handler
- Network-policy reviews verifying that internal listeners are not broadly reachable
- Policy-as-code checks requiring authenticated service-to-service communication and least-privilege permissions
- Runtime monitoring for unauthenticated access attempts, unexpected proxy operations, command execution, and secret reads
Automated security review tools may help flag exposed sensitive routes, missing middleware, and overly broad service permissions.

