Currently, MCP connectors configured with static-header authentication for use in workflow AI nodes are also exposed in Port Chat and the Port MCP server, with no way to isolate them to a specific invocation context. Would like a scope setting on MCP connector entities that allows admins to restrict a connector's availability to either workflow (workflow AI nodes only), chat (Port Chat / Port MCP server only), or all (current default behavior). This would prevent workflow-only connectors from being unintentionally surfaced to end users in interactive chat, and allow teams to maintain a clean separation between automation-focused and user-facing MCP connectors.