uv-mcp
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill describes tools like
repair_environment,add_dependency, andbuild_projectwhich perform system-level operations to manage virtual environments and packages. These are standard and expected capabilities for a project management skill. - [EXTERNAL_DOWNLOADS]: The skill mentions capabilities for installing the
uvbinary and various Python interpreters. These operations involve downloading artifacts from external sources, which is the primary purpose of the described tools. - [PROMPT_INJECTION]: The instructions focus purely on operational guidance for the uv-mcp server. No patterns attempting to bypass safety filters or override agent behavior were identified.
- [DATA_EXFILTRATION]: There are no patterns indicating the unauthorized access or external transmission of sensitive information such as credentials or environment variables.
Audit Metadata