CVE-2026-55689: OpenFGA OIDC Authentication Bypass via Missing Audience Validation
OpenFGA, an authorization engine widely used by developers, has a flaw in how it validates authentication tokens when using OpenID Connect (OIDC). Before version 1.18.0, if a system administrator configured OIDC authentication without explicitly setting an audience parameter, the system would accept tokens meant for other services using the same identity provider. An attacker with valid credentials for a related service could potentially use their token to gain unauthorized access to OpenFGA instances. This is a configuration-dependent vulnerability that requires both OIDC enablement and a missing audience configuration to be exploitable.
Source data · NVD / CISA · public domain
- CVSS
- 3.1 · 6.8 MEDIUM · CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:N
- Weaknesses (CWE)
- CWE-287
- Affected products
- 2 configuration(s)
- Published / Modified
- 2026-07-09 / 2026-07-14
NVD description (verbatim)
OpenFGA is an authorization/permission engine built for developers. Prior to 1.18.0, OpenFGA's OIDC authenticator skipped JWT audience validation when authn.method was set to oidc, authn.oidc.issuer was configured, and authn.oidc.audience was not set, allowing a token minted for an unrelated service by the same identity provider to authenticate to OpenFGA. This issue is fixed in 1.18.0.
5 reference(s) · View on NVD →
SEC.co analysis · AI-assisted, reviewed against source
Technical summary
CVE-2026-55689 is an authentication bypass vulnerability stemming from incomplete JWT validation in OpenFGA's OIDC authenticator. When authn.method is set to 'oidc' and authn.oidc.issuer is configured, the system skipped the JWT audience (aud) claim validation if authn.oidc.audience was not explicitly set. This allows tokens issued by the same identity provider but intended for unrelated services to successfully authenticate against the affected OpenFGA instance. The vulnerability is rooted in CWE-287 (Improper Authentication) and affects the authorization decision chain at the entry point. Version 1.18.0 introduces proper audience validation, making the aud claim check mandatory when OIDC is enabled.
Business impact
Organizations relying on OpenFGA for access control decisions face unauthorized access risk if they have deployed vulnerable versions with OIDC authentication enabled. The impact depends on what protected resources or operations the OpenFGA instance governs. An attacker with valid credentials in the same identity provider ecosystem (e.g., a former contractor with access to a related internal service) could bypass authentication controls and potentially escalate privileges or access sensitive data. The vulnerability is most dangerous in multi-tenant or microservices environments where token reuse across service boundaries is possible. Remediation is straightforward and does not require architectural changes.
Affected systems
OpenFGA versions prior to 1.18.0 are affected when deployed with OIDC authentication configured. Both the openfga/openfga container image and openfga Helm charts delivering versions before 1.18.0 are in scope. Systems using alternative authentication methods (API keys, mTLS, or local authentication) are not affected. Environments where authn.oidc.audience is already explicitly configured are protected from this specific attack vector.
Exploitability
Exploitation requires an attacker to hold valid credentials issued by the same OIDC identity provider and have network access to the OpenFGA instance. The CVSS score of 6.8 (MEDIUM) reflects the combination of high impact (confidentiality and integrity) offset by the higher barriers to entry (authentication required, less common configuration). The vulnerability is not known to be actively exploited in the wild and has not been added to CISA's Known Exploited Vulnerabilities catalog. An attacker cannot exploit this remotely without valid OIDC credentials, which is the primary limiting factor.
Remediation
Upgrade OpenFGA to version 1.18.0 or later. For organizations unable to upgrade immediately, implement defense-in-depth by explicitly configuring the authn.oidc.audience parameter to restrict accepted tokens to those minted specifically for your OpenFGA deployment. Additionally, conduct an audit of identity provider trust relationships and enforce principle of least privilege within OIDC issuer configurations. Monitor authentication logs for tokens originating from unexpected service identities.
Patch guidance
Obtain OpenFGA 1.18.0 or later from the official openfga/openfga repository. If deploying via Helm, update the chart to a version that bundles 1.18.0 or later of the OpenFGA application. Verify the patch is in place by confirming the application version in startup logs or by checking the /health endpoint if exposed. Test authentication flows in a non-production environment before production rollout to ensure OIDC configuration remains compatible with your identity provider setup.
Detection guidance
Search logs for authentication requests using OIDC tokens where the audience claim does not match your configured service identity. Enable verbose logging on the OIDC authenticator if available. Examine OIDC token metadata (via JWT decoding) for evidence of tokens minted for unrelated services. Monitor for failed authorization decisions following successful authentication, which may indicate token reuse from unintended sources. If possible, enable audit logging on your identity provider to detect tokens issued for your OpenFGA service being used elsewhere.
Why prioritize this
This vulnerability warrants prompt but not emergency attention. The CVSS score of 6.8 and authentication requirement mean it is less critical than network-reachable unauthenticated flaws, but the direct impact on authorization decisions makes it more serious than many medium-severity issues. Organizations with OIDC-based OpenFGA deployments should prioritize this within a standard patching cycle. The fix is low-risk and the configuration gap is easily identified through review of existing deployment manifests.
Risk score, explained
The CVSS 3.1 score of 6.8 (MEDIUM) reflects: high impact on confidentiality and integrity (C:H/I:H) because successful exploitation bypasses access controls; network accessibility (AV:N); requirement for prior authentication (PR:L), which limits attacker population; and high complexity (AC:H) due to the specific configuration required to trigger the flaw. The score appropriately captures that this is not a critical flaw but poses meaningful risk in certain deployment patterns.
Frequently asked questions
Do we need to upgrade if we're not using OIDC authentication?
No. This vulnerability only affects OpenFGA instances with authn.method set to 'oidc' and authn.oidc.issuer configured. If you authenticate via API keys, mutual TLS, or other methods, you are not exposed to this flaw.
We already set authn.oidc.audience in our configuration. Are we vulnerable?
No. The vulnerability exists only when authn.oidc.audience is NOT set. If you have explicitly configured this parameter, JWT audience validation is enforced and this specific bypass cannot occur. Verify your running configuration to confirm the value is present.
Can this be exploited without network access to OpenFGA?
No. The vulnerability requires an attacker to send an HTTP request to OpenFGA with a valid OIDC token. An attacker must have both valid credentials from the shared identity provider and network connectivity to your OpenFGA endpoint.
What should we do while waiting to patch?
As an immediate measure, ensure authn.oidc.audience is explicitly configured in your OpenFGA deployment to match your service identity. Review OIDC token logs for anomalies. Expedite the upgrade to 1.18.0 within your standard patching schedule; the fix is low-risk and requires no configuration changes beyond the upgrade itself.
This analysis is provided for informational purposes and based on the CVE record published 2026-07-09. It does not constitute legal, compliance, or professional security advice. Verify all version numbers, patch availability, and vendor guidance against official sources before making deployment decisions. Organizations should conduct their own risk assessment based on their specific infrastructure and threat model. SEC.co assumes no liability for implementation decisions made in reliance on this analysis. Source: NVD (public-domain), retrieved 2026-08-17. Analysis generated by SEC.co (claude-haiku-4-5).
Related vulnerabilities
- CVE-2023-5502MEDIUMArista EOS 802.1x Authentication Bypass Vulnerability
- CVE-2026-10283MEDIUMBottelet DaybydayCRM Authentication Bypass in Settings Handler
- CVE-2026-10548MEDIUMImproper Authentication in NousResearch hermes-agent Credential Synchronization
- CVE-2026-13208MEDIUMKubeVirt virt-handler gRPC Identity Validation Flaw
- CVE-2026-13543MEDIUMDocumenso Google OAuth Authentication Bypass Vulnerability (CVSS 5.6)
- CVE-2026-14627MEDIUMNousResearch Hermes-Agent Discord Authentication Bypass Vulnerability
- CVE-2026-14714MEDIUMCowAgent 2.1.0 WeChat Authentication Bypass – Exploit, Patch & Mitigation
- CVE-2026-15192MEDIUMMissing Authentication in Mettle Sendportal APIv1 Webhooks