GHSA-gppg-gqw8-wh9g
litellm vulnerable to remote code execution based on using eval unsafely
Packages
Details
As written in the advisory (Markdown source, shown as plain text).
BerriAI/litellm version v1.35.8 contains a vulnerability where an attacker can achieve remote code execution. The vulnerability exists in the `add_deployment` function, which decodes and decrypts environment variables from base64 and assigns them to `os.environ`. An attacker can exploit this by sending a malicious payload to the `/config/update` endpoint, which is then processed and executed by the server when the `get_secret` function is triggered. This requires the server to use Google KMS and a database to store a model.
References
External links from the advisory. They lead to third-party sites.