The Agent Skills Directory

[SAFE]: The skill facilitates the orchestration of Model Context Protocol (MCP) servers, providing guidance on discovery via standard tools like grok inspect and identifying capability gaps in the current environment.
[SAFE]: The instructions include a dedicated Safety Rules section that explicitly warns against installing untrusted or overly broad MCP servers without review, particularly those that can execute arbitrary code or access secrets.
[SAFE]: The skill interacts with standard agent configuration files, such as .grok/config.toml, as part of its documented workflow for tool installation and integration.
[SAFE]: Recommended tools (e.g., agent-session-analyzer, repo-graph, test-intelligence) are aligned with the skill's stated purpose and intended for development and analysis tasks.

mcp-orchestrator