test-mcp-server
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the use of the
ngrokCLI to manage tunnels, secrets, and cloud endpoints (e.g.,ngrok api,ngrok http). These commands are standard for the vendor's intended functionality of providing secure remote access to local services. - [INDIRECT_PROMPT_INJECTION]: The skill creates a network-accessible surface that processes requests from external AI providers, which is an inherent property of the MCP protocol implementation.
- Ingestion points: External data enters the local environment through the
/mcpPOST endpoint defined inreferences/http-transport.ts. - Boundary markers: The
traffic-policy.yamlconfiguration acts as a security boundary by enforcing Bearer token authentication and path-based filtering before traffic reaches the local server. - Capability inventory: The skill utilizes
expressfor local server hosting and thengrokagent to establish a secure tunnel. - Sanitization: Communication is handled through the official
@modelcontextprotocol/sdk, which provides the necessary protocol framing and handling.
Audit Metadata