The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/evaluation.py and scripts/connections.py files utilize the Model Context Protocol SDK to execute local shell commands (via stdio_client) to launch and interact with MCP servers. This is a core functional requirement for the skill's primary purpose of building and testing server implementations.
[INDIRECT_PROMPT_INJECTION]: The evaluation framework in scripts/evaluation.py creates a potential attack surface by ingesting test questions from an external XML file (eval_file) and processing them through an LLM that has tool execution capabilities. While the skill uses XML tags in the EVALUATION_PROMPT as boundary markers to structure the agent's output, the processing of untrusted input questions represents a known vulnerability surface common in automated testing workflows.

mcp-builder