Supply-chain attack on LiteLLM stole cloud credentials via malicious packages
The claim
The authors assert that attackers compromised the LiteLLM open-source AI gateway to steal cloud credentials and test stolen AWS keys.
Published as: “26th October – Threat Intelligence Report”
So what
Compromised AI gateway packages can give attackers direct access to cloud accounts, expanding the impact of supply-chain attacks.
The five-minute version
The report describes a supply-chain compromise of the open-source AI gateway LiteLLM on October 21, where attackers used stolen developer tokens to publish malicious packages and Docker images (version 1.107.0). The malicious release was designed to exfiltrate cloud credentials and AI service secrets, and the attackers later tested the stolen AWS credentials. The evidence is based on the vendor’s confirmation and observed malicious package distribution, indicating a real-world breach.
What they found 2
- Attackers used stolen developer tokens to publish malicious LiteLLM packages and Docker images.
“LiteLLM, a widely used open-source AI gateway, confirmed an October 21 supply-chain attack in which attackers used stolen developer tokens to publish malicious packages and Docker images.”
- The compromised LiteLLM version was designed to steal cloud credentials and AI service secrets, and attackers tested stolen AWS credentials.
“The compromised version 1.107.0 was designed to steal cloud credentials and AI service secrets, with attackers subsequently testing stolen AWS credentials.”
What’s new here 3
- First public disclosure of a LiteLLM supply-chain compromise using stolen developer tokens.
- Identification of malicious package version 1.107.0 that harvests cloud credentials.
- Evidence that attackers actively tested stolen AWS keys after deployment.
Artifacts & methods 3
- LiteLLMsystem
- malicious LiteLLM package version 1.107.0primitive
- stolen developer tokensprimitive