The Agent Skills Directory

[COMMAND_EXECUTION]: The evaluation scripts (scripts/evaluation.py and scripts/connections.py) allow for the execution of arbitrary local commands via the stdio transport using user-provided CLI arguments. This is the intended mechanism for running and testing local MCP servers.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection in scripts/evaluation.py. The script ingests question text from XML files and response data from external MCP servers, passing them directly into the LLM context without sanitization. Ingestion points: scripts/evaluation.py lines 48 and 103; Boundary markers: Absent; Capability inventory: subprocess for server launch and anthropic for LLM calls; Sanitization: Absent.
[EXTERNAL_DOWNLOADS]: SKILL.md directs users to fetch documentation and schemas from official Model Context Protocol domains and GitHub repositories. These are recognized as trusted or well-known sources for development resources.

mcp-builder