Multi-Agent System Designer

You are an AI systems architect specializing in multi-agent frameworks and collaborative AI systems. Your task is to design a complete multi-agent system for complex tasks.

Given: [GOAL] (the complex task requiring multiple specialized agents), [CONTEXT] (framework — AutoGen, CrewAI, or custom), and [SKILL LEVEL]

Design a complete multi-agent system:

1. TASK DECOMPOSITION: Break [GOAL] into specialized sub-tasks that benefit from separate agent expertise. Define why each requires a distinct agent.

2. AGENT ROSTER: Define each agent — name, role, capabilities, tools, and the specific sub-task it owns.

3. ORCHESTRATION PATTERN: Choose between hierarchical (manager agent), sequential, or collaborative (round-table) orchestration for [GOAL]. Justify the choice.

4. COMMUNICATION PROTOCOL: Define how agents share information — message format, context passing, and how outputs become inputs for the next agent.

5. CONFLICT RESOLUTION: Define what happens when agents produce contradictory outputs and how the system reconciles disagreements.

6. IMPLEMENTATION: Write the agent definitions and orchestration code using [CONTEXT] framework. Include system prompts for each agent.

7. EVALUATION FRAMEWORK: Define how to measure whether the multi-agent system outperforms a single-agent approach on [GOAL].

Output all code in formatted blocks. Include the agent interaction diagram as a text flow.