CVE-2026-31942: LibreChat IDOR Vulnerability Allows Unauthorized API Key Manipulation

Technical summary

CVE-2026-31942 is an Insecure Direct Object Reference (IDOR) vulnerability in LibreChat's API key management endpoint (PUT /api/keys). The vulnerability stems from unsafe use of the JavaScript object spread operator following authenticated user ID assignment. When a request is processed, the code sets the user's ID, then applies a spread operator that merges additional properties from the request body. An attacker can exploit this by sending a crafted request with a userId parameter to override another user's configuration. The vulnerability affects the storage and routing of API credentials for third-party AI providers. It is patched in version 0.8.3-rc1.

Business impact

If exploited, this vulnerability could result in unauthorized access to or manipulation of API credentials, leading to several business risks: stolen API keys could be monetized or used to exhaust quota limits; conversations could be rerouted through attacker-controlled infrastructure, creating data exfiltration and compliance risks; service denial attacks become possible by replacing keys with invalid credentials; and customers may face unexpected billing or service interruptions from unauthorized API consumption. For SaaS or self-hosted LibreChat deployments serving multiple users, the reputational and financial impact could be significant.

Affected systems

LibreChat versions 0.7.6 and earlier are vulnerable. The issue has been resolved in version 0.8.3-rc1. Organizations running self-hosted LibreChat instances or relying on LibreChat as part of their AI infrastructure stack should audit their deployment versions immediately.

Exploitability

Exploitation requires valid authentication to LibreChat—an attacker must have a user account. Once authenticated, no special privileges or user interaction is required; the attack is carried out via a simple API request with an injected userId parameter. The network attack surface is accessible, making this feasible for any authenticated user. The CVSS score of 7.1 (HIGH) reflects the combination of low authentication requirements and significant integrity and availability impacts.

Remediation

Upgrade to version 0.8.3-rc1 or later as soon as possible. The patch corrects the object spread operator logic to prevent userId injection. Organizations should also audit existing API key configurations to identify any unauthorized changes or suspicious key history logs. If available, enable API key usage monitoring and rate-limit alerts to detect anomalous consumption patterns.

Patch guidance

Verify your current LibreChat version and upgrade to 0.8.3-rc1 or any patched release version beyond that. If you are running a stable release version, check the official LibreChat repository or vendor advisories for the corresponding patched release. Test the upgrade in a non-production environment first to ensure compatibility with your deployment. Note that version 0.8.3-rc1 is a release candidate; confirm against the official vendor advisory for the recommended production-grade patch version.

Detection guidance

Monitor API request logs for PUT requests to /api/keys containing userId parameters in the request body that differ from the authenticated user's own ID. Check for suspicious changes to API key configurations by reviewing audit logs or key modification timestamps. Look for requests from users attempting to modify keys belonging to other users. Implement alerts on failed API key validation attempts, which may indicate an attacker testing replaced credentials. If available, enable verbose logging of API key assignments and scope changes.

Why prioritize this

This vulnerability deserves immediate attention because it requires only user-level authentication to execute, affects all users in a multi-user LibreChat deployment, and enables both confidentiality (via credential theft) and availability (via service denial) attacks. The high CVSS score and the direct control over API credentials create a significant risk posture. Multi-tenant deployments are at particular risk.

Risk score, explained

The CVSS 3.1 score of 7.1 (HIGH) is driven by: low attack complexity (L), requiring only low privilege (L) user authentication (PR:L), no user interaction (UI:N), unchanged scope (S:U), no confidentiality impact from the vulnerability itself (C:N), but high integrity impact (I:H) due to the ability to modify any user's API keys, and low availability impact (A:L) from the potential for service disruption. The combination of easy exploitation post-authentication and significant integrity compromise justifies the HIGH severity rating.

Frequently asked questions