The Agent Skills Directory

[SAFE]: The skill provides structured guidance for implementing MCP servers. All provided instructions and patterns align with the official Model Context Protocol specifications and best practices.
[COMMAND_EXECUTION]: The scripts/connections.py file uses the official mcp library to manage subprocesses via the stdio transport layer. This is the standard method for local MCP server integration and is used appropriately within the context of the skill's primary purpose.
[EXTERNAL_DOWNLOADS]: Documentation references point to verified resources on modelcontextprotocol.io and official GitHub repositories under the modelcontextprotocol organization. These are trusted sources for protocol specifications and SDKs.
[SAFE]: The evaluation script in scripts/evaluation.py uses the Anthropic SDK to test MCP tools. It handles API keys through standard environment variables and includes boundary markers in its prompts to ensure structured AI responses, following secure development principles.

mcp-builder