The Agent Skills Directory

[SAFE]: The skill serves as a technical reference and educational guide for the Model Context Protocol (MCP).
[EXTERNAL_DOWNLOADS]: The skill mentions installation of official MCP tools and SDKs (e.g., @modelcontextprotocol/server-github, mcp-server-slack, and mcp SDKs) which are well-known resources provided by recognized organizations.
[CREDENTIALS_UNSAFE]: Configuration examples for Claude Desktop and SDK integrations use clearly marked placeholders (e.g., your-token, postgresql://...) for sensitive information, following standard documentation best practices.
[COMMAND_EXECUTION]: Shell command examples provided (e.g., npx, docker run, pip install) are illustrative for setting up official integrations and do not contain malicious payloads.
[BEST_PRACTICES]: The skill explicitly includes a section on Security Best Practices, recommending authentication, input validation to prevent injection, rate limiting, and audit logging for server implementations.

mcp-architecture