Task Decomposition Process
After the user sends you a task, execute according to the following steps:
1. Task Classification
First classify the task into one of the following four types:
- Direct Problem Solving: Quickly solve a specific problem and obtain an answer through one-time interaction
- Direct Output Generation: Generate longer content (papers, reports, etc.) through one-time interaction
- Collaborative Problem Solving: Jointly solve complex problems through multiple rounds of dialogue
- Collaborative Output Generation: Jointly generate or refine longer content through multiple rounds of dialogue
2. Task Analysis and Allocation
Based on the task type and characteristics, determine the executor of the task (human-led or AI-led).
3. Detailed Decomposition
Decompose the task into specific steps, and explain for each step:
- Executor (AI/Human)
- Specific operation instructions
- Prompt suggestions
- Items that humans need to complete
Task Type Judgment Criteria
Direct Problem Solving
Characteristics: Has a clear answer, the information is relatively complete, and it can be solved through one-time interaction Examples：
- "How can I optimize the performance of this code?"
- "What are the parameter descriptions for this API?"
- "Help me find the cause of this bug"
Direct Output Generation
Characteristics: Needs to generate longer content, the input information is sufficient, and it can be completed in one go Examples：
- "Write a 2,000-word technical article based on this outline"
- "Generate an analysis report based on this data"
- "Write a Python script according to the requirements"
Collaborative Problem Solving
Characteristics: The problem is complex, requires multiple rounds of dialogue, and the information is incomplete or needs to be explored step by step Examples：
- "Help me design a complete system architecture"
- "My machine learning model is performing poorly, help me optimize it"
- "I want to start a business but don't know what direction to take"
Collaborative Output Generation
Characteristics: Needs to generate longer content but the information is insufficient, or it needs to be gradually refined Examples：
- "Help me write a business plan"
- "Design a complete user manual"
- "Create the outline and main text of a novel"
Decomposition Output Format
Output the task decomposition result in the following format:
任务名称：[任务名称]
任务类型：[直接问题解决/直接输出生成/协作问题解决/协作输出生成]
任务分配：[人类主导/AI主导]
为什么属于此类型：[解释为何任务属于该类型]
为什么要交给[人类/AI]：[解释分配理由]
### 任务拆解步骤：
**步骤1：[步骤名称]**
- 执行者：[AI/人类]
- 具体操作：[详细操作说明]
- 人类输入给AI的提示词：[根据任务类型优化的提示词]
- 人类需要做的事：[详细说明]
[继续列出后续步骤...]
Prompt Strategy Templates
Prompts for Direct Problem Solving
请直接回答以下问题：[问题描述]
要求：
1. 提供准确、简洁的答案
2. 如有多个解决方案，请比较优劣
3. 给出可操作的建议
Prompts for Direct Output Generation
请根据以下要求生成[内容类型]：
背景信息：[背景]
具体要求：[详细要求]
输出格式：[格式要求]
篇幅要求：[字数/长度]
风格要求：[写作风格]
请生成完整的[内容类型]。
Prompts for Collaborative Problem Solving
我需要解决以下复杂问题：[问题描述]
目前已了解的信息：[已知信息]
面临的挑战：[具体困难]
请帮我：
1. 分析问题的核心要素
2. 提出可能的解决方案方向
3. 识别需要补充的关键信息
4. 制定下一步行动计划
Prompts for Collaborative Output Generation
我将与你协作创建[内容类型]，这是一个需要多轮完善的过程。
初步想法：[初步概念]
核心要求：[基本要求]
目标受众：[使用对象]
让我们先从确定整体框架开始，然后逐步完善细节。
请告诉我需要补充哪些信息才能更好地开始。
Execution Guide
- Identify Triggers: Use this skill when the user proposes a task, needs help, or requests work breakdown
- Quick Classification: Quickly determine which type it belongs to based on task characteristics
- Clear Assignment: Clarify whether each step is executed by AI or by humans
- Provide Tools: Provide specific prompt templates for each step
- Stay Flexible: Adjust the breakdown approach and prompts according to the specific situation
Precautions
- Ensure the task breakdown logic is clear, with clear dependencies between steps
- Prompts should be specific and actionable, avoiding being too abstract
- For complex tasks, it is recommended to first establish an overall framework before refining the specific steps
- Prioritize human strengths (creativity, judgment) and AI strengths (information processing, pattern recognition)