The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/evaluation.py and scripts/connections.py files include functionality to execute local commands and arguments to launch MCP servers using the stdio transport. This is achieved through the mcp library's stdio_client, which manages subprocess execution. This is a core feature for developers to test and debug their MCP server implementations.
[PROMPT_INJECTION]: The skill implements an evaluation harness that processes questions from XML files and integrates outputs from external MCP servers into the agent's context. This setup presents a surface for indirect prompt injection if the data being processed contains malicious instructions. However, this is a known characteristic of agentic workflows and is mitigated here by the focused, developer-centric purpose of the tools.
[EXTERNAL_DOWNLOADS]: The SKILL.md file contains instructions to fetch official documentation and SDK information from the modelcontextprotocol GitHub organization. These are well-known, trusted sources in the MCP ecosystem and are documented neutrally as resources for the user.

mcp-builder