building-mcp-servers
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (LOW): The skill instructs the agent to execute a local script
scripts/verify.pyand the commandnpx @modelcontextprotocol/inspector. These are standard operations for verifying the development environment and testing server tools. - [EXTERNAL_DOWNLOADS] (INFO): The skill references SDK README files and protocol documentation from
githubusercontent.comandmodelcontextprotocol.io. These are legitimate external resources required for implementing the server. - [PROMPT_INJECTION] (LOW): Indirect Prompt Injection surface (Category 8). Ingestion point: Fetching documentation from
modelcontextprotocol.io. Boundary markers: None. Capability inventory: Command execution (npm,npx,python). Sanitization: None. The threat is considered low as the data is used to inform the agent's reasoning for the specific development task without direct side effects from the data itself.
Audit Metadata