Temporal Server compiles a Worker Controller Instance module, go.temporal.io/auto-scaled-workers, into its Worker Service, and that module registers a compute provider named k8s whose function is to manage workers running in a Kubernetes cluster. The k8s provider built its Kubernetes client from the client configuration supplied in the caller's request rather than from operator configuration, and building and using that client could run a command of the caller's choosing on the Worker Service host, under the account the server process runs as. Two frontend operations reach the provider. CreateWorkerDeploymentVersion requires a namespace write role. ValidateWorkerDeploymentVersionComputeConfig is registered as a read-only namespace operation, so an authenticated caller holding only a read role in a single namespace can reach the provider through it, and that route does not require a worker deployment to already exist. Execution happens while the submitted configuration is validated, so no scaling decision, task arrival, or unusual request sequence is required. Because the Worker Service process holds the persistence credentials for every namespace in the cluster and the cluster's TLS material, the consequence reaches beyond the caller's namespace to the cluster as a whole. The provider is present in the official temporal-server binaries and container images for the affected releases. The control that keeps it unreachable is the compute provider allowlist, the dynamic configuration setting workercontroller.compute_providers.enabled. On Temporal Server 1.31.0 through 1.31.2 that allowlist does not deny by default: when the setting is unset the check is skipped and every registered compute provider is permitted, k8s included, so a stock deployment of those releases is exposed. On 1.31.3 and later 1.31 releases, and on 1.32.0 and later 1.32 releases, the allowlist denies by default, and a deployment is exposed only where k8s has been added to it. The 1.30 release line does not include the Worker Controller Instance module and is unaffected. To determine whether a deployment is affected, check the following together. The deployed Temporal Server version is in an affected range. The Worker Service is running, which it is in the default service set and therefore in a stock deployment. The effective value of workercontroller.compute_providers.enabled contains k8s, or, on 1.31.0 through 1.31.2 only, is unset; check both the global value and every namespace-constrained override, because an override can add k8s for a single namespace. And authorization is configured, meaning a real authorizer and claim mapper are in place; a deployment running with no authorizer already grants every caller unrestricted access to every namespace, so it has no namespace boundary for this to cross. Deployments that use a custom authorizer should also check whether it permits ValidateWorkerDeploymentVersionComputeConfig to callers without namespace write access. The separate per-namespace dynamic configuration setting workercontroller.enabled does not gate the affected path on any affected release.
" } ] } ], "references": [ { "url": "https://github.com/temporalio/temporal/releases/tag/v1.32.2", "name": "First fixed 1.32 release", "tags": [ "release-notes" ] }, { "url": "https://github.com/temporalio/temporal/releases/tag/v1.31.5", "name": "First fixed 1.31 release", "tags": [ "release-notes" ] }, { "url": "https://github.com/temporalio/temporal-auto-scaled-workers/pull/165", "name": "Fix pull request, Kubernetes compute provider removed from default builds", "tags": [ "patch" ] }, { "url": "https://github.com/temporalio/temporal/pull/12843", "name": "Fix pull request, compute configuration validation requires namespace write access", "tags": [ "patch" ] }, { "url": "https://www.cve.org/CVERecord?id=CVE-2026-89139", "name": "CVE-2026-89139, related compute provider vulnerability in the same module" } ], "metrics": [ { "format": "CVSS", "scenarios": [ { "lang": "en", "value": "GENERAL" } ], "cvssV4_0": { "attackVector": "NETWORK", "attackComplexity": "LOW", "attackRequirements": "NONE", "privilegesRequired": "LOW", "userInteraction": "NONE", "vulnConfidentialityImpact": "HIGH", "subConfidentialityImpact": "NONE", "vulnIntegrityImpact": "HIGH", "subIntegrityImpact": "NONE", "vulnAvailabilityImpact": "HIGH", "subAvailabilityImpact": "NONE", "exploitMaturity": "NOT_DEFINED", "Safety": "NOT_DEFINED", "Automatable": "NOT_DEFINED", "Recovery": "NOT_DEFINED", "valueDensity": "NOT_DEFINED", "vulnerabilityResponseEffort": "NOT_DEFINED", "providerUrgency": "NOT_DEFINED", "version": "4.0", "baseSeverity": "HIGH", "baseScore": 8.7, "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N" } } ], "configurations": [ { "lang": "en", "value": "Exploitation requires all of the following at once. The deployment runs an affected Temporal Server release with the Worker Service enabled, which is the default service set, so a stock deployment satisfies this. The k8s compute provider is permitted by the dynamic configuration setting workercontroller.compute_providers.enabled, either globally or for the caller's namespace. On 1.31.0 through 1.31.2 an unset value permits every registered provider, so no operator action is needed there; on 1.31.3 and later 1.31 releases and on 1.32 releases, the operator must have added k8s to the list. Authorization is configured through an authorizer and a claim mapper; without it every caller already holds unrestricted access to every namespace and there is no namespace boundary to cross. The caller holds a read role in some namespace, which is what the frontend requires to validate a worker deployment version's compute configuration, or a write role, which is what it requires to create a worker deployment version. No other feature flag has to be enabled by the operator: system.enableDeploymentVersions defaults to true, and the per-namespace Worker Controller component does not depend on workercontroller.enabled on any affected release.", "supportingMedia": [ { "type": "text/html", "base64": false, "value": "Exploitation requires all of the following at once. The deployment runs an affected Temporal Server release with the Worker Service enabled, which is the default service set, so a stock deployment satisfies this. The k8s compute provider is permitted by the dynamic configuration setting workercontroller.compute_providers.enabled, either globally or for the caller's namespace. On 1.31.0 through 1.31.2 an unset value permits every registered provider, so no operator action is needed there; on 1.31.3 and later 1.31 releases and on 1.32 releases, the operator must have added k8s to the list. Authorization is configured through an authorizer and a claim mapper; without it every caller already holds unrestricted access to every namespace and there is no namespace boundary to cross. The caller holds a read role in some namespace, which is what the frontend requires to validate a worker deployment version's compute configuration, or a write role, which is what it requires to create a worker deployment version. No other feature flag has to be enabled by the operator: system.enableDeploymentVersions defaults to true, and the per-namespace Worker Controller component does not depend on workercontroller.enabled on any affected release.
" } ] } ], "workarounds": [ { "lang": "en", "value": "Remove k8s from the dynamic configuration setting workercontroller.compute_providers.enabled, at the global level and in every namespace-constrained override. On Temporal Server 1.31.0 through 1.31.2, also set the setting to an explicit list naming only the compute providers the deployment intends to permit, omitting k8s, because leaving it unset permits every registered provider; an explicit empty list denies every compute provider. The request is then rejected with InvalidArgument before the provider runs. This is dynamic configuration, so it takes effect without restarting the server and is available during an incident rather than only at a maintenance window. Deployments that use a custom authorizer can additionally require namespace write access for ValidateWorkerDeploymentVersionComputeConfig, which removes the read-role route but not the write-role route.", "supportingMedia": [ { "type": "text/html", "base64": false, "value": "Remove k8s from the dynamic configuration setting workercontroller.compute_providers.enabled, at the global level and in every namespace-constrained override. On Temporal Server 1.31.0 through 1.31.2, also set the setting to an explicit list naming only the compute providers the deployment intends to permit, omitting k8s, because leaving it unset permits every registered provider; an explicit empty list denies every compute provider. The request is then rejected with InvalidArgument before the provider runs. This is dynamic configuration, so it takes effect without restarting the server and is available during an incident rather than only at a maintenance window. Deployments that use a custom authorizer can additionally require namespace write access for ValidateWorkerDeploymentVersionComputeConfig, which removes the read-role route but not the write-role route.
" } ] } ], "solutions": [ { "lang": "en", "value": "Upgrade to Temporal Server 1.32.2 or 1.31.5. Both remove the k8s compute provider from default builds, so it is no longer compiled into the official binaries and container images, and both require a namespace write role for ValidateWorkerDeploymentVersionComputeConfig. Operators who cannot upgrade immediately can apply the workaround below, which closes the same path without waiting for a release. The 1.30 release line does not depend on the Worker Controller Instance module at all and is unaffected. Note that the fixed releases remove the provider from default builds and do not change where its Kubernetes client configuration comes from: an operator who builds Temporal Server with the k8s compute provider included and permits it in the allowlist still accepts that configuration from the caller's request.", "supportingMedia": [ { "type": "text/html", "base64": false, "value": "Upgrade to Temporal Server 1.32.2 or 1.31.5. Both remove the k8s compute provider from default builds, so it is no longer compiled into the official binaries and container images, and both require a namespace write role for ValidateWorkerDeploymentVersionComputeConfig. Operators who cannot upgrade immediately can apply the workaround below, which closes the same path without waiting for a release. The 1.30 release line does not depend on the Worker Controller Instance module at all and is unaffected. Note that the fixed releases remove the provider from default builds and do not change where its Kubernetes client configuration comes from: an operator who builds Temporal Server with the k8s compute provider included and permits it in the allowlist still accepts that configuration from the caller's request.
" } ] } ], "credits": [ { "lang": "en", "value": "An external security researcher who reported this issue responsibly to Temporal Technologies", "type": "finder" } ], "source": { "discovery": "EXTERNAL" } } } }