Dynamic Multi-Action AI Tool Nodes (Unified Tool Definition for AI Agents)

Feature Description Currently, when connecting service nodes (such as Gmail, Slack, Notion, GitHub) as Tools to an AI Agent node, each tool node instance is pinned to a single deterministic Resource + Operation (e.g., Message -> Send or Message -> Get Many). If an AI Agent needs full interaction with a service (e.g. searching, reading, drafting, and sending emails), workflow builders must drag 4–5 separate Tool nodes onto the canvas, wire each one individually into the AI Agent, configure credentials repeatedly, and manually maintain tool descriptions. In contrast, modern agentic tool ecosystems (e.g., Model Context Protocol / MCP, Composio, OpenClaw, Hermes) expose a service under a unified dynamic tool definition where the LLM selects the operation and parameters at runtime based on natural language instructions. Proposed Improvement Introduce a "Dynamic / Multi-Action" execution mode on AI Tool nodes: 1. Dynamic Tool Mode on Node-as-a-Tool: - In the tool no…