Update LangChat Now to Block Unauthenticated API Key Theft and Super-Admin Creation
A critical flaw in LangChat allows remote attackers to create super-admin accounts and steal provider API keys without any credentials.
Key points
- CVE-2026-111704 is rated Critical with a CVSS score of 9.3 due to missing authentication on REST controllers.
- Attackers can retrieve plaintext passwords and LLM provider API keys by querying the /api/v1/auth/users endpoint.
- The vulnerability exists in LangChat versions through commit 2a6285a because no Sa-Token interceptor is registered.
Remote attackers can fully compromise LangChat installations by accessing sensitive data and administrative controls without providing any credentials. The flaw, identified as CVE-2026-111704, grants unauthenticated users the ability to create super-admin accounts, modify or delete existing data, and read plaintext passwords. This exposure also includes the theft of large language model provider API keys.
According to the National Vulnerability Database, the issue stems from a missing authentication vulnerability in LangChat through commit 2a6285a. The absence of a registered Sa-Token interceptor allows attackers to bypass security checks entirely. By querying the /api/v1/auth/users endpoint, malicious actors can retrieve critical credentials and assume full control over the application. The National Vulnerability Database rates this issue as Critical with a CVSS score of 9.3, reflecting the severe impact of unauthenticated remote access.
Context and Background
The National Vulnerability Database classifies this weakness under CWE-306, which relates to missing authentication for a critical function. In this specific instance, the failure to register a Sa-Token interceptor leaves REST controllers exposed to public internet traffic. This architectural oversight means that any request to protected endpoints is processed as if it came from an authorized user. Consequently, the application fails to verify the identity of the requester before granting access to sensitive resources.
Who Is Affected
Any organization running LangChat versions up to and including commit 2a6285a is vulnerable to this attack. The vulnerability allows complete system takeover, including the creation of persistent administrative accounts. Because the attack requires no initial credentials, remote attackers can access the system directly. The National Vulnerability Database notes that attackers can query specific endpoints to read plaintext passwords and retrieve API keys.
What to watch
No patch or fix has been confirmed yet by the vendor. Users should monitor official LangChat channels for an update that addresses CVE-2026-111704. Until a fix is available, organizations should consider restricting external access to the application. Security teams should also audit logs for unauthorized access to the /api/v1/auth/users endpoint.
What made it possible
Missing authentication vulnerabilities occur when applications fail to verify user identity before granting access to sensitive functions. This weakness typically arises from misconfigured security interceptors or omitted validation steps in API endpoints. Attackers exploit this gap to perform actions reserved for authorized users, such as data retrieval or account creation. Teams usually mitigate this risk by ensuring every protected endpoint enforces strict authentication checks. Regular audits of API security configurations help identify these gaps before they are exploited.
What to do and how to stay safe: LangChat
- Restrict network access to the LangChat application to trusted internal IP addresses only.
- Monitor application logs for unauthorized requests to the /api/v1/auth/users endpoint.
- Rotate all LLM provider API keys and user passwords immediately if exposure is suspected.
- Wait for an official vendor update before applying any unofficial patches.
Step-by-step guide: How Security Advisories Drive Patch Decisions and Reduce Risk
General security guidance from the Shell Bulletin newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is the CVSS score for CVE-2026-111704?
The National Vulnerability Database rates CVE-2026-111704 with a CVSS score of 9.3, classifying it as Critical.
Which versions of LangChat are affected by this vulnerability?
LangChat versions through commit 2a6285a are affected by this missing authentication vulnerability.
Can attackers access API keys through this flaw?
Yes, unauthenticated remote attackers can retrieve LLM provider API keys by querying the /api/v1/auth/users endpoint.
