mcp-builder
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- COMMAND_EXECUTION (LOW): The
MCPConnectionStdioclass inscripts/connections.pyenables the execution of system commands via the MCP stdio transport. This is a core protocol feature but requires the execution environment to be secured.\n- EXTERNAL_DOWNLOADS (LOW): TheMCPConnectionSSEandMCPConnectionHTTPclasses inscripts/connections.pyallow connections to external URLs, which could be used for data retrieval or interaction with remote services.\n- PROMPT_INJECTION (LOW): The skill is susceptible to indirect prompt injection via tool outputs.\n - Ingestion points: Data returned by the
call_toolmethod inscripts/connections.pyis integrated into the agent's context.\n - Boundary markers: No boundary markers or instruction-ignoring delimiters are implemented for tool results.\n
- Capability inventory: The skill can execute local commands (
stdio) and initiate network requests (SSE/HTTP).\n - Sanitization: There is no evidence of sanitization or validation of the content received from MCP servers before it is returned to the agent.
Audit Metadata