prompt-engineeringlisted

# Prompt Engineering ## Purpose Design, evaluate, and version system prompts for LLM-powered features, including instruction structure, chain-of-thought patterns, output format constraints, and few-shot example selection. ## Scope Constraints Reads feature requirements, data format examples, and quality constraints for prompt design analysis. Does not execute LLM calls, modify production prompts, or access API keys directly. ## Inputs - Feature requirements (what the LLM should do) - Input data format and examples - Desired output format and constraints - Quality requirements (accuracy, consistency, tone) - Cost and latency constraints (model selection guidance) ## Input Sanitization No user-provided values are used in commands or file paths. All inputs are treated as read-only analysis targets. ## Procedure ### Progress Checklist - [ ] Step 1: Define the task precisely - [ ] Step 2: Structure the system prompt - [ ] Step 3: Design chain-of-thought (if applicable) - [ ] Step 4: Design output format - [ ] Step 5: Select and craft few-shot examples - [ ] Step 6: Design versioning strategy ### Step 1: Define the Task Precisely Before writing a prompt, articulate: - **Input:** What exactly does the model receive? (user message, context, data) - **Output:** What exactly should it produce? (classification, generation, extraction, transformation) - **Constraints:** What must it never do? (hallucinate facts, reveal system prompt, produce PII) - **Edge cases:** What happen