返回提示词库
文本 · 通用大模型OpenAI Agents SDK 客服系统多Agent编排提示词PW
创作者Prompt2 编辑部PromptWhisper 收录
文本通用大模型开发与工程

OpenAI Agents SDK 客服系统多Agent编排提示词

使用 OpenAI Agents SDK 设计多Agent协作的智能客服系统,包含路由、专业处理和升级机制

13浏览
完整提示词可替换花括号中的变量后直接使用

You are a system architect designing a multi-agent customer service system using the OpenAI Agents SDK. Generate a complete agent orchestration plan. ## System Requirements Business Type: [BUSINESS_TYPE] Support Channels: [chat/email/voice] Languages: [LANGUAGES] ## Agent Architecture ### 1. Triage Agent (Router) Role: Classify incoming requests and route to specialist agents Tools: classify_intent, check_customer_tier, fetch_context Handoff Rules: - billing_issues -> Billing Agent - technical_support -> Tech Agent - general_inquiry -> FAQ Agent - escalation_needed -> Human Escalation Agent ### 2. Specialist Agent Template For each specialist agent, define: - System Prompt: Role description with domain expertise - Tools: List of function tools (API calls, DB queries, actions) - Guardrails: Input/output validators to prevent hallucination - Handoff Conditions: When to transfer to another agent or escalate - Context Window: What customer data to inject ### 3. Output Format For each agent in the system, provide a JSON object with: agent_name, role, system_prompt, tools (array of name/description/parameters), guardrails (input/output arrays), handoff_rules (condition/target_agent pairs), and max_turns. ### 4. Conversation Flow Map the complete conversation flow as a state diagram, including: - Entry points per channel - Decision nodes for routing - Parallel processing opportunities - Escalation paths - Resolution confirmation loops Generate the complete multi-agent system design for my business.

2026/4/27

如何使用这条提示词

  1. 1复制上方完整提示词。
  2. 2在对应模型中替换主题、人物或风格变量。
  3. 3生成后记录有效调整,形成自己的版本。