Skip to content

Security/Dependency Risk: Update litellm constraint to prevent pulling compromised versions #146

@dustarion

Description

@dustarion

Description

I noticed in the memgraph-toolbox/pyproject.toml that the toolkit specifies litellm>=1.77.1.

"litellm>=1.77.1",

The open-ended >= constraint puts users at risk of automatically downloading compromised packages during a pip install. Recently, litellm was targeted in a supply chain attack on PyPI where malicious versions (1.82.7 and 1.82.8) containing credential stealers were published.

Recommendation

To protect users of the Memgraph AI Toolkit from this and future supply chain attacks, I highly recommend pinning this dependency strictly to the last known safe version or capping the upper limit.

Steps to Reproduce

Check the current dependency declarations for litellm.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions