CVE-2026-31944: LibreChat
CVE-2026-31944
LibreChat is a ChatGPT clone with additional features. From 0.8.2 to 0.8.2-rc3, The MCP (Model Context Protocol) OAuth callback endpoint accepts the redirect from the identity provider and stores OAut...
Overview
A significant security vulnerability, tracked as CVE-2026-31944, has been identified in LibreChat, an open-source ChatGPT clone. This flaw affects versions 0.8.2 through 0.8.2-rc3. It is an OAuth authentication bypass that could allow an attacker to hijack a victim’s linked external service accounts.
Vulnerability Explained
In simple terms, this vulnerability exists in the feature that allows LibreChat to connect to external services like Atlassian or Outlook using OAuth (a common “Log in with…” protocol). The flaw is in the callback process. When a user completes an OAuth login, LibreChat incorrectly stores the received access tokens without verifying who is currently logged into the LibreChat application itself.
An attacker can exploit this by creating a malicious OAuth authorization link for a service and sending it to a victim. If the victim clicks the link and completes the login for that external service, the victim’s access tokens for that service (e.g., their Jira or Microsoft 365 credentials) are automatically and silently stored in the attacker’s LibreChat account. The attacker then gains full control over the victim’s account on that linked external service.
Impact and Risk
The impact of this vulnerability is severe, rated as HIGH with a CVSS score of 7.6. Successful exploitation leads directly to account takeover (ATO) of any MCP-linked service the victim authorizes. This could result in:
- Unauthorized access to corporate Atlassian (Jira, Confluence), Google, Microsoft 365, or other integrated accounts.
- Data theft, data manipulation, or further lateral movement within an organization’s ecosystem.
- Significant privacy and security breaches stemming from compromised credentials.
For context on the dangers of account takeover, recent incidents are often detailed in public breach reports.
Remediation and Mitigation
The primary and essential remediation is to update LibreChat immediately.
Action Required:
- Upgrade: All users must upgrade to LibreChat version 0.8.3-rc1 or later, where this vulnerability has been patched. No workarounds are provided; upgrading is the only complete fix.
- Investigate: Administrators of affected instances should review audit logs for any suspicious OAuth authorization events prior to the patch and instruct users to review connected accounts and revoke any that are unfamiliar. Users of compromised external services should reset their passwords and review account activity.
- General Security Hygiene: This flaw underscores the importance of scrutinizing authentication flows. Organizations should implement regular software update cycles and stay informed on security news for similar vulnerabilities in other platforms.
If you are running an affected version, treat this update with high priority to prevent potential account compromise.
Never miss a critical vulnerability
Get real-time security alerts delivered to your preferred platform.
Related Advisories
C4G Basic Laboratory Information System 3.4 contains multiple SQL injection vulnerabilities that allow unauthenticated attackers to execute arbitrary SQL commands by injecting malicious code through t...
VPN Browser+ 1.1.0.0 contains a denial of service vulnerability that allows unauthenticated attackers to crash the application by submitting oversized input through the search functionality. Attackers...
PraisonAI is a multi-agent teams system. In versions below 4.5.139 of PraisonAI and 1.5.140 of praisonaiagents, the browser bridge (praisonai browser start) is vulnerable to unauthenticated remote ses...
Missing authentication for critical function in Azure MCP Server allows an unauthorized attacker to disclose information over a network....