Horizon Alert
Summary of the vulnerability and why it matters
This critical vulnerability in LiteLLM, an AI gateway, allows an unauthenticated attacker to read or modify data in the proxy's database. This could lead to unauthorized access to sensitive credentials managed by the proxy.
- Attacker can access sensitive credentials.
- Affects systems using LiteLLM as a proxy.
- Requires no special access to exploit.
Attack Path
How an attacker could exploit the issue
An unauthenticated attacker can exploit this vulnerability by sending a malicious Authorization header to the LiteLLM proxy's API endpoints. This crafted header triggers an error-handling path, leading to a SQL injection that can expose or modify sensitive data within the proxy's database. This grants unauthorized access to the proxy and any managed credentials.
- No authentication required.
- Target API routes with Authorization header.
- Exploits error handling for SQL injection.
Live Threat
Current exploitation, exposure, and threat context
This SQL injection vulnerability in LiteLLM is attractive to attackers because it allows them to directly query the proxy's database, potentially exposing sensitive credentials and internal data. The vulnerability is present in a core function of the proxy, making it broadly impactful if exploited.
- Listed on KEV.
- Public exploit likely.
- Exploited in the wild.
Priority actions
Operational Fix
Recommended remediation, mitigation, and detection steps
Prioritize updating LiteLLM to version 1.83.7 to address the critical SQL injection vulnerability. If immediate patching is not feasible, isolate affected services to prevent exploitation of the database access vulnerability.
- Update LiteLLM to 1.83.7.
- Isolate affected services.
- Monitor for suspicious database access.
