CVE-2026-31866: Allocation of Resources Without Limits or Throttling in flagd
Details
flagd exposes OFREP (/ofrep/v1/evaluate/...) and gRPC (evaluation.v1, evaluation.v2) endpoints for feature flag evaluation. These endpoints are designed to be publicly accessible by client applications.
The evaluation context included in request payloads is read into memory without any size restriction. An attacker can send a single HTTP request with an arbitrarily large body, causing flagd to allocate a corresponding amount of memory. This leads to immediate memory exhaustion and process termination (e.g., OOMKill in Kubernetes environments).
flagd does not natively enforce authentication on its evaluation endpoints. While operators may deploy flagd behind an authenticating reverse proxy or similar infrastructure, the endpoints themselves impose no access control by default.
Impact
- Denial of Service: A single crafted request can crash the flagd process. - Service Disruption: All applications relying on the affected flagd instance for feature flag evaluation will lose access to flag evaluations until the process restarts. - Repeated Exploitation: An attacker can continuously send oversized requests to prevent recovery.
Affected Endpoints
- /ofrep/v1/evaluate/flags/{flagKey} (OFREP single flag evaluation) - /ofrep/v1/evaluate/flags (OFREP bulk evaluation) - flagd.evaluation.v1.Service/ResolveBoolean (gRPC/Connect) - flagd.evaluation.v1.Service/ResolveString (gRPC/Connect) - flagd.evaluation.v1.Service/ResolveFloat (gRPC/Connect) - flagd.evaluation.v1.Service/ResolveInt (gRPC/Connect) - flagd.evaluation.v1.Service/ResolveObject (gRPC/Connect) - flagd.evaluation.v1.Service/ResolveAll (gRPC/Connect) - flagd.evaluation.v2.Service/ResolveBoolean (gRPC/Connect) - flagd.evaluation.v2.Service/ResolveString (gRPC/Connect) - flagd.evaluation.v2.Service/ResolveFloat (gRPC/Connect) - flagd.evaluation.v2.Service/ResolveInt (gRPC/Connect) - flagd.evaluation.v2.Service/ResolveObject (gRPC/Connect)
Other sources
flagd is a feature flag daemon with a Unix philosophy. Prior to 0.14.2, flagd exposes OFREP (/ofrep/v1/evaluate/...) and gRPC (evaluation.v1, evaluation.v2) endpoints for feature flag evaluation. These endpoints are designed to be publicly accessible by client applications. The evaluation context included in request payloads is read into memory without any size restriction. An attacker can send a single HTTP request with an arbitrarily large body, causing flagd to allocate a corresponding amount of memory. This leads to immediate memory exhaustion and process termination (e.g., OOMKill in Kubernetes environments). flagd does not natively enforce authentication on its evaluation endpoints. While operators may deploy flagd behind an authenticating reverse proxy or similar infrastructure, the endpoints themselves impose no access control by default. This vulnerability is fixed in 0.14.2.
— MITRE
Affected Software
Remediation
Event History
Frequently Asked Questions
What is the severity of CVE-2026-31866?
CVE-2026-31866 has been classified as a high-severity vulnerability due to its potential to expose sensitive data.
How do I fix CVE-2026-31866?
To fix CVE-2026-31866, upgrade to version 0.14.3 or later of the flagd package.
What software is affected by CVE-2026-31866?
CVE-2026-31866 affects the flagd package version 0.14.2 and earlier.
What are the consequences of exploiting CVE-2026-31866?
Exploitation of CVE-2026-31866 may allow unauthorized access to feature flag evaluation endpoints and sensitive information.
Is there a workaround for CVE-2026-31866?
A potential workaround for CVE-2026-31866 is to restrict access to the affected endpoints until the software is updated.