uv-advanced
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- EXTERNAL_DOWNLOADS (HIGH): The documentation in
references/docker.mdprovides an installation pattern that downloads a shell script fromhttps://astral.sh/uv/install.sh. While this is the official source for the tool, the domainastral.shis not included in the explicitly trusted source list, making the download and subsequent execution a high-risk pattern under an 'assume-malicious' posture. - REMOTE_CODE_EXECUTION (HIGH): Several reference files (
references/projects.md,references/pip-interface.md) describe workflows for adding and installing packages from arbitrary remote Git repositories and URLs (e.g.,uv add git+https://github.com/org/repoanduv pip install https://...). This allows for the installation and execution of unvetted remote code within the agent's environment. - COMMAND_EXECUTION (MEDIUM): The skill documentation promotes the use of
uv runanduvx(ephemeral tool execution) to run arbitrary Python packages and CLI tools. These commands represent a significant execution surface, as they can download and run code from external sources in a single step.
Recommendations
- AI detected serious security threats
Audit Metadata