The Agent Skills Directory

[SAFE]: The skill fetches documentation and project details from official Model Context Protocol domains and GitHub repositories, which are well-known and trusted sources.- [COMMAND_EXECUTION]: The evaluation script (scripts/evaluation.py) is designed to launch and interact with MCP servers. When using the stdio transport, it executes a command provided via arguments to start the server process, which is the standard mechanism for local MCP integration.- [PROMPT_INJECTION]: The evaluation harness processes input from XML files to test LLM responses. This ingestion surface is a necessary part of the testing functionality. The evaluator agent is provided with a specific system prompt to guide its behavior during these tests.

mcp-builder