building-mcp-servers
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references and fetches documentation from official and trusted sources, including the Model Context Protocol website and official GitHub repositories for the MCP SDKs.
- [COMMAND_EXECUTION]: The evaluation scripts (
scripts/evaluation.pyandscripts/connections.py) allow for the execution of local commands and arguments. This functionality is intended to allow developers to launch and test their own MCP servers during the development process. - [INDIRECT_PROMPT_INJECTION]: The evaluation harness parses user-provided XML files to extract test questions which are then processed by an LLM. While this presents a surface for indirect prompt injection, it is the primary mechanism for the tool's intended purpose of testing LLM behavior against an MCP server.
Audit Metadata