CVE-2025-64751: OpenFGA Helm Charts Auth Bypass Flaw

CVE-2025-64751 Overview

CVE-2025-64751 affects OpenFGA, a high-performance authorization and permission engine inspired by Google Zanzibar. The flaw allows improper policy enforcement when specific Check and ListObject calls are executed against vulnerable deployments. Affected versions include OpenFGA Docker images v1.4.0 through v1.11.0 and Helm charts openfga-0.1.34 through openfga-0.2.48. The vulnerability is categorized under CWE-285: Improper Authorization and was patched in version 1.11.1.

Critical Impact
Authenticated callers can receive incorrect authorization decisions from OpenFGA, leading to access being granted or denied contrary to the configured authorization model.

Affected Products

OpenFGA server v1.4.0 through v1.11.0 (Docker)
OpenFGA Helm chart openfga-0.1.34 through openfga-0.2.48
Applications relying on OpenFGA Check and ListObject APIs for access decisions

Discovery Timeline

2025-11-21 - CVE-2025-64751 published to NVD
2025-12-31 - Last updated in NVD database

Technical Details for CVE-2025-64751

Vulnerability Analysis

OpenFGA implements a relationship-based access control (ReBAC) model in which authorization decisions are computed by evaluating relations between users, objects, and types. Client applications query the engine through Check calls, which return a boolean authorization decision, and ListObject calls, which enumerate the objects a subject can access for a given relation.

In the affected releases, certain combinations of authorization model definitions and query inputs cause the policy evaluator to return incorrect results. The defect breaks the contract that the engine enforces the authorization model defined by the application owner, producing decisions inconsistent with the intended policy.

The issue is classified as an improper authorization weakness [CWE-285]. Exploitation requires network access to the OpenFGA API and valid low-privilege credentials, but does not require user interaction. The vulnerability impacts subsequent systems that consume OpenFGA decisions, not the OpenFGA process itself.

Root Cause

The root cause is a defect in the policy evaluation logic that handles specific Check and ListObject query paths. The evaluator fails to apply the configured authorization model correctly under those conditions, producing answers that diverge from the model semantics. Refer to the GitHub Security Advisory GHSA-2c64-vmv2-hgfc for the specific technical conditions identified by the maintainers.

Attack Vector

An attacker with API access to a vulnerable OpenFGA instance issues crafted Check or ListObject requests that trigger the faulty evaluation path. Because the OpenFGA decision is consumed by downstream services as the source of truth for authorization, an incorrect allow decision propagates as a broken access control condition in the calling application. The attack vector is network-based and requires only low-privilege caller permissions to the OpenFGA API.

No verified public proof-of-concept is available. Technical specifics are described in the upstream v1.11.1 release notes.

Detection Methods for CVE-2025-64751

Indicators of Compromise

OpenFGA server logs showing unexpected allow responses for Check calls that contradict the deployed authorization model.
ListObject responses returning objects that the requesting subject should not be able to enumerate.
Application audit logs recording resource access by principals whose direct or inherited relations do not justify the access.

Detection Strategies

Inventory all OpenFGA deployments and compare the running version against the vulnerable ranges v1.4.0 to v1.11.0 and Helm chart openfga-0.1.34 to openfga-0.2.48.
Replay representative Check and ListObject queries against a patched v1.11.1 instance and diff the responses to surface decisions affected by the defect.
Correlate OpenFGA API access logs with downstream application authorization events to identify discrepancies between expected and actual access outcomes.

Monitoring Recommendations

Enable verbose OpenFGA request logging and forward decisions to a centralized log store for retention and analysis.
Alert on anomalous spikes in Check allow rates or ListObject cardinality for sensitive object types.
Track the OpenFGA container image digest and Helm chart version in CI/CD pipelines and trigger alerts when a vulnerable version is deployed.

How to Mitigate CVE-2025-64751

Immediate Actions Required

Upgrade OpenFGA to version 1.11.1 or later across all environments, including Docker images and Helm chart deployments.
Restrict network exposure of the OpenFGA API to trusted service identities until the upgrade is complete.
Audit recent authorization decisions for sensitive object types to confirm no unintended access occurred while running an affected version.

Patch Information

The maintainers fixed the issue in OpenFGA v1.11.1. Container images and updated Helm charts are available from the OpenFGA v1.11.1 release. Review the GitHub Security Advisory GHSA-2c64-vmv2-hgfc for advisory metadata and version constraints.

Workarounds

No vendor-supplied workaround is published. Upgrading to v1.11.1 is the supported remediation.
Where immediate upgrade is not feasible, enforce a secondary authorization check in the calling application for high-risk operations until the OpenFGA upgrade is rolled out.
Limit which client identities can call the Check and ListObject endpoints using network policy or API gateway controls.

bash

# Upgrade the OpenFGA Helm release to a fixed chart version
helm repo update openfga
helm upgrade openfga openfga/openfga --version 0.2.49 \
  --set image.tag=v1.11.1

# Verify the running server version
kubectl exec deploy/openfga -- /openfga version