JackaL
友一人
- Joined
- Sep 3, 2026
- Messages
- 341
- Reaction score
- 61
[GUIDE] Master Prompting: Context-Action-Format (CAF) Framework
Stop getting generic AI responses. Whether you are using ChatGPT or the Claude API, applying the CAF framework guarantees high-quality, production-ready outputs.
1. Context (The Persona)
Define exactly who the AI is and what constraints exist. Example: "Act as a senior Python developer specializing in async scraping."
2. Action (The Task)
Be extremely specific about the operation. Avoid vague words like "write" or "make." Use "refactor," "summarize," or "generate a script that..."
3. Format (The Output)
Tell the LLM exactly how you want the data presented to avoid parsing errors in your applications.
Stop getting generic AI responses. Whether you are using ChatGPT or the Claude API, applying the CAF framework guarantees high-quality, production-ready outputs.
1. Context (The Persona)
Define exactly who the AI is and what constraints exist. Example: "Act as a senior Python developer specializing in async scraping."
2. Action (The Task)
Be extremely specific about the operation. Avoid vague words like "write" or "make." Use "refactor," "summarize," or "generate a script that..."
3. Format (The Output)
Tell the LLM exactly how you want the data presented to avoid parsing errors in your applications.
Code:
Output requirements:
- Use markdown bullet points.
- Provide the code in a single block.
- Do not include explanatory text before the code.