Skip to main content
Vulnerability Database/CVE-2025-46732

CVE-2025-46732: OpenCTI Authentication Bypass Vulnerability

CVE-2025-46732 is an authentication bypass flaw in OpenCTI that allows authenticated users to access, modify, and delete other users' notifications via IDOR. This post explains its impact, affected versions, and mitigation steps.

Published:

CVE-2025-46732 Overview

CVE-2025-46732 is an Insecure Direct Object Reference (IDOR) vulnerability in OpenCTI, an open source cyber threat intelligence platform. The flaw resides in the GraphQL NotificationLineNotificationMarkReadMutation and NotificationLineNotificationDeleteMutation mutations. Authenticated users who know the UUID of another user's notification can read, modify the read status of, or delete that notification. When flipping the read status, the attacker also receives the notification's content, exposing sensitive intelligence data. Versions prior to 6.6.6 are affected. The vulnerability is categorized under CWE-285: Improper Authorization.

Critical Impact

Any authenticated OpenCTI user can disclose, tamper with, or destroy notifications belonging to other users, including administrators, breaking confidentiality and integrity of threat intel workflows.

Affected Products

  • OpenCTI Platform versions prior to 6.6.6
  • Citeum OpenCTI deployments exposing the GraphQL notification API
  • Any OpenCTI tenant with multiple authenticated users

Discovery Timeline

  • 2025-07-18 - CVE-2025-46732 published to the National Vulnerability Database (NVD)
  • 2026-06-17 - Last updated in NVD database

Technical Details for CVE-2025-46732

Vulnerability Analysis

OpenCTI exposes a GraphQL API for user-facing operations, including notification management. Two mutations handle notification lifecycle actions: NotificationLineNotificationMarkReadMutation toggles the read flag, and NotificationLineNotificationDeleteMutation removes a notification.

Both mutations accept a notification UUID as input. The resolvers verify that the caller is authenticated but fail to verify that the caller owns the referenced notification. As a result, any logged-in user can operate on any notification by supplying its identifier.

The mark-read mutation additionally returns the notification payload in its response. Attackers therefore gain read access to the notification body, not merely the ability to mutate state. This exposes the targeted user's threat intel alerts, mentions, and digest content.

Root Cause

The root cause is a missing authorization check in the notification resolvers. Per CWE-285: Improper Authorization, the application confirms authentication but does not confirm that the authenticated principal is the notification owner before executing the mutation.

Attack Vector

Exploitation requires network access to the OpenCTI GraphQL endpoint and valid credentials for any role. The attacker must obtain or guess a target notification UUID. UUIDs may leak through shared links, logs, API responses in multi-tenant workflows, or lower-privileged inspection of related objects. Once the UUID is known, the attacker issues a single GraphQL mutation request. See the OpenCTI GitHub Security Advisory GHSA-535g-qp2c-h7vp for the authoritative description.

Detect Identity Attacks the Moment Authentication Fails

Singularity™ Identity surfaces credential misuse, lateral movement, and privilege escalation in real time, so you can contain the breach before it spreads.

Stop Identity Attacks

Detection Methods for CVE-2025-46732

Indicators of Compromise

  • GraphQL requests invoking NotificationLineNotificationMarkReadMutation or NotificationLineNotificationDeleteMutation from a user whose session does not own the referenced notification UUID
  • Unexpected deletions of notifications reported by end users or administrators
  • Notifications flipping between read and unread states without user interaction
  • Access logs showing a single account issuing notification mutations across many distinct UUIDs in a short time window

Detection Strategies

  • Correlate the authenticated user ID in GraphQL request logs with the owner ID of the notification UUID referenced in each mutation; flag mismatches
  • Alert on bursts of markRead or delete notification mutations from a single session, which indicate UUID enumeration or scripted abuse
  • Review audit logs for notification state changes initiated by non-owner principals

Monitoring Recommendations

  • Enable verbose GraphQL request logging in OpenCTI, including operation name, variables, and caller identity
  • Forward OpenCTI application and reverse proxy logs to a SIEM or data lake for retention and correlation
  • Monitor the OpenCTI version banner to confirm all instances are running 6.6.6 or later

How to Mitigate CVE-2025-46732

Immediate Actions Required

  • Upgrade OpenCTI to version 6.6.6 or later, which contains the fix for the notification authorization checks
  • Audit GraphQL access logs for prior invocations of the two notification mutations and identify cross-user access
  • Restrict GraphQL endpoint exposure to trusted networks and enforce strong authentication for all OpenCTI accounts
  • Rotate API tokens for accounts suspected of misuse during the exposure window

Patch Information

The OpenCTI maintainers released version 6.6.6, which adds ownership verification to NotificationLineNotificationMarkReadMutation and NotificationLineNotificationDeleteMutation. Refer to the OpenCTI GitHub Security Advisory GHSA-535g-qp2c-h7vp for release details and upgrade instructions.

Workarounds

  • No official workaround is published; patching to 6.6.6 is the supported remediation
  • As interim hardening, limit OpenCTI accounts to trusted personnel and review user roles to reduce the attacker population
  • Place the OpenCTI GraphQL endpoint behind an authenticating reverse proxy and apply rate limiting to slow UUID enumeration

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.