The Agent Skills Directory

[SAFE]: The skill is entirely educational and documentation-focused, providing guidance on how to use the Model Context Protocol (MCP) securely.
[COMMAND_EXECUTION]: The documentation describes the use of stdio MCP servers, which involve executing local child processes. It provides examples using npx, python, and local script paths, which are standard for this protocol.
[EXTERNAL_DOWNLOADS]: Includes examples of running packages from the official @modelcontextprotocol organization using npx. These are recognized as well-known and trusted technology sources.
[PROMPT_INJECTION]: The skill details how to build agents that ingest data from external services via MCP tools. While this architectural pattern creates a surface for indirect prompt injection, the skill includes mitigation advice such as using granular tool permissions and avoiding wildcards.

MCP Integration