Cursor Experiments

Demo: Multi-Agent Orchestration in Cursor

This project demonstrates how specialized AI agents can collaborate to solve complex, multi-step tasks within the Cursor IDE.

Project Goal

The primary goal is to showcase the **orchestration capability** of Cursor's agentic framework. By defining a high-level "Trip Planner" agent that coordinates with specialized "Flight" and "Hotel" researchers, we illustrate a modular and scalable approach to automated problem-solving.


Key Features

• Multi-Agent Collaboration: Transparent delegation of sub-tasks from a primary agent to specialized sub-agents.
• Skill-Based Execution: Integration of custom "Agent Skills" that execute shell scripts to perform specific actions (e.g., searching data).

Subagents and Agent Skills

Agents

• @trip-planner: Orchestrates the entire trip planning process by delegating tasks to specialized agents.
• @flight-researcher: Researches flight options based on origin and destination.
• @hotel-researcher: Researches hotel accommodations at the destination.

Agent Skills

• @research-flight: Executes flight search scripts to retrieve travel options.
• @research-hotel: Executes hotel search scripts to retrieve accommodation details.

Relationships across subagents and Agent Skills

graph TD
    TripPlanner[Trip Planner Agent] --> FlightResearcher[Flight Researcher Agent]
    TripPlanner --> HotelResearcher[Hotel Researcher Agent]
    FlightResearcher --> ResearchFlight[Research Flight Skill]
    HotelResearcher --> ResearchHotel[Research Hotel Skill]

Usage Instructions

To run the demonstration:

1. Open the Cursor IDE in this repository.
2. Open the Chat panel (Cmd+L or Ctrl+L).
3. Use the @trip-planner agent by typin