uv
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides standard documentation for the 'uv' tool, which is a legitimate and widely-used Python package manager.
- [EXTERNAL_DOWNLOADS]: The skill references downloads from well-known and official sources, including the GitHub Container Registry ('ghcr.io/astral-sh/uv'), the official Astral 'setup-uv' GitHub Action, and the official PyTorch package index. These sources are considered reputable.
- [COMMAND_EXECUTION]: The skill facilitates the execution of Python scripts and CLI tools within project environments using 'uv run' and 'uvx'. This functionality is standard for a package management tool.
- [SAFE]: Indirect Prompt Injection Surface (Category 8): The skill presents an attack surface by processing project configuration files and executing associated commands.
- Ingestion points: The skill reads and processes 'pyproject.toml', 'uv.lock', and environment files ('.env') within project directories.
- Boundary markers: Boundary markers are not explicitly defined in the instructions, as the skill relies on the 'uv' tool's native handling of configuration files.
- Capability inventory: The tool possesses capabilities for file system writes (virtual environment creation), network operations (dependency synchronization), and command execution via 'uv run' and 'uvx'.
- Sanitization: Relies on the 'uv' tool's internal validation of project metadata and lockfiles.
- Note: This surface is a standard characteristic of package management tools and is not a malicious pattern.
Audit Metadata