The Agent Skills Directory

[COMMAND_EXECUTION]: The evaluation harness (evaluation.py) and connection manager (connections.py) facilitate the execution of local shell commands and scripts. This is the intended functionality for launching and testing MCP servers as part of the development lifecycle.
[EXTERNAL_DOWNLOADS]: The skill instructions guide the agent to fetch protocol specifications and SDK documentation from official sources, including modelcontextprotocol.io and GitHub. These are well-known resources for the Model Context Protocol.
[INDIRECT_PROMPT_INJECTION]: The evaluation script processes external data from XML files and tool outputs. While this creates an attack surface where malicious data could influence agent behavior, the risk is inherent to testing harnesses and is mitigated by the use of XML boundary markers in the system prompt and the intended use case for local developer testing.

mcp-builder