Impact
The vulnerability is located in the get_secret function of the secret resolution component in LiteLLM. When an attacker supplies a crafted api_key parameter, the function bypasses its intended authorization checks and returns secrets that should be protected. This flaw allows the attacker to read sensitive credentials or configuration data that an authenticated user should not see, thereby compromising confidentiality of secrets within the system. This flaw maps to CWE-639: Abuse of Authorization.
Affected Systems
The issue is present in BerriAI LiteLLM releases up to and including version 1.94.0. It affects the secret_managers/main.py module in the Secret Resolution component. Versions newer than 1.94.0 are presumed to not contain the flaw, but the vendor did not confirm a fix.
Risk and Exploitability
The CVSS score of 5.3 indicates a medium severity vulnerability. EPSS is not available, so the likelihood of exploitation is unclear. The flaw can be triggered remotely through the get_secret API, and the vendor has not released a patch or response. The vulnerability is not listed in the CISA KEV catalog. Given the potential for unauthorized secret disclosure, the risk is significant and exploitation could lead to credential compromise or further lateral movement within the environment.
OpenCVE Enrichment