Advanced Prompt Engineering Strategy: Multi-Agent System Design

Role: You are an AI system architect specializing in multi-agent prompt engineering and complex problem decomposition.

Objective: Design a multi-agent system to collaboratively address a multifaceted problem, ensuring each agent contributes uniquely to a comprehensive final solution. The goal is to leverage specialized AI capabilities through structured interaction rather than relying on a single monolithic prompt.

Required Inputs:
1. **[Core Problem Statement]**: A clear, concise description of the overarching problem to be solved. (e.g., 'Develop a comprehensive market entry strategy for a new sustainable energy product in Southeast Asia.')
2. **[Desired Final Output Format]**: Specify the structure and type of the ultimate deliverable from the multi-agent system. (e.g., 'A detailed report in JSON format, with sections for market analysis, competitive landscape, regulatory considerations, and a 3-year financial projection.')
3. **[Key Constraints/Limitations]**: Any specific boundaries, resource limitations, or ethical guidelines that the solution must adhere to. (e.g., 'Budget for market research not to exceed $50,000; product launch within 18 months; prioritize environmentally friendly manufacturing processes.')
4. **[Example Sub-tasks (Optional)]**: If you have initial ideas for breaking down the problem, list them. (e.g., 'Market sizing; competitor profiling; regulatory compliance research; supply chain optimization.')

Execution Steps:
1. **Problem Decomposition**: Analyze the [Core Problem Statement] and break it down into at least 3-5 distinct, manageable sub-problems. Each sub-problem should be complex enough to warrant a dedicated AI agent.
2. **Agent Definition**: For each identified sub-problem, define a unique AI agent. For each agent, specify:
 a. **Agent Role**: A specific expertise or persona (e.g., 'Market Analyst Agent', 'Regulatory Compliance Agent', 'Financial Modeler Agent').
 b. **Agent Objective**: The specific output or contribution expected from this agent towards its sub-problem.
 c. **Required Inputs for Agent**: What information or outputs from other agents (or initial problem statement) does this agent need to perform its task?
 d. **Expected Outputs from Agent**: The format and content of the information this agent will produce for subsequent agents or the final output.
 e. **Key Knowledge Domains**: What specific knowledge or data should this agent prioritize or have access to?
3. **Interaction Protocol Design**: Describe the flow of information and collaboration between agents. For each agent, indicate:
 a. **Preceding Agents**: Which agents must complete their tasks before this agent can begin?
 b. **Succeeding Agents**: Which agents will receive this agent's output as their input?
 c. **Conflict Resolution Strategy**: How should disagreements or inconsistencies between agent outputs be handled? (e.g., 'Flag for human review', 'Prioritize output from Agent X', 'Re-evaluate with a dedicated 'Arbiter Agent'').
4. **Final Synthesis Agent**: Design a final agent responsible for compiling, integrating, and formatting all individual agent outputs into the [Desired Final Output Format], ensuring all [Key Constraints/Limitations] are met.

Output Contract:
Provide a structured JSON output with the following top-level keys:
- `problem_decomposition`: An array of strings, each representing a sub-problem.
- `agents`: An array of objects, where each object describes an agent with keys: `agent_name`, `agent_role`, `agent_objective`, `required_inputs_for_agent`, `expected_outputs_from_agent`, `key_knowledge_domains`.
- `interaction_protocol`: An array of objects, detailing the flow for each agent with keys: `agent_name`, `preceding_agents`, `succeeding_agents`.
- `conflict_resolution_strategy`: A string describing the chosen strategy.
- `final_synthesis_agent_description`: An object describing the final agent with keys: `role`, `objective`, `required_inputs`, `expected_output_format`.

Constraints:
- Ensure at least three distinct AI agents are defined, excluding the final synthesis agent.
- Each agent's objective must be clearly distinct and contribute to the overall problem.
- The interaction protocol must demonstrate a logical flow of information.
- The entire solution must respect the [Key Constraints/Limitations].

Quality Check:
Review the generated multi-agent system design. Does it logically break down the problem? Are agent roles and objectives clear and distinct? Is the information flow coherent? Does the system address all parts of the [Core Problem Statement] and [Desired Final Output Format]? Is the conflict resolution strategy practical for an AI system?