The Agent Skills Directory

[COMMAND_EXECUTION]: The evaluation harness (scripts/evaluation.py) and connection utility (scripts/connections.py) facilitate the execution of local commands through the stdio transport. The script takes user-supplied commands and arguments to launch MCP server processes for testing purposes. While functional, this provides the agent with a mechanism to spawn local subprocesses.
[EXTERNAL_DOWNLOADS]: The skill instructs agents to fetch protocol specifications and SDK documentation from official sources, including modelcontextprotocol.io and the modelcontextprotocol organization on GitHub. These references are used to ensure the agent has up-to-date context on the MCP standard and are documented neutrally as they originate from well-known repositories.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to build servers that ingest and process data from external services, which introduces a surface for indirect prompt injection via tool outputs or resource content. The provided evaluation harness also processes XML files that could contain malicious instructions. The skill proactively addresses this by including an ASSESSMENT.md and QA.md that guide agents to implement boundary markers, strict schema validation, and input sanitization to mitigate these risks.

mcp-builder