mcp-server
Fail
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches the installation script for the 'uv' package manager from the official Astral domain (astral.sh) as part of the recommended developer setup.- [REMOTE_CODE_EXECUTION]: Executes the downloaded 'uv' installation script via a shell pipe. This is the standard installation method for this developer tool and is included in the project setup instructions.- [INDIRECT_PROMPT_INJECTION]: The skill provides templates for building tools that ingest and process external API data, which introduces a potential surface for indirect prompt injection if the source data is compromised.
- Ingestion points: Tools like 'get_alerts' and 'get_forecast' in SKILL.md accept user-provided parameters (state, latitude, longitude) to query the National Weather Service API.
- Boundary markers: The provided example code does not implement specific delimiters or 'ignore' instructions to separate API responses from the agent's instructions.
- Capability inventory: The server templates use 'httpx' for network operations and the MCP SDK for tool execution and transport.
- Sanitization: The examples do not demonstrate sanitization or validation of the content returned from external APIs before it is passed back to the agent context.
Recommendations
- HIGH: Downloads and executes remote code from: https://astral.sh/uv/install.sh - DO NOT USE without thorough review
Audit Metadata