What problem does it solve? Writing prompts that LLMs reliably follow is hard: vague instructions produce inconsistent outputs, unstructured system prompts drift, and untested changes silently degrade quality. This Skill treats prompt engineering like software engineering, providing repeatable structure, patterns, and QA processes. ## Core Features & Use Cases - Structured Prompt Architecture: A standard system prompt pattern (Role → Context → Instructions → Constraints → Examples → Output Format) for consistent, maintainable prompts. - Advanced Techniques: Few-shot learning, chain-of-thought reasoning, progressive disclosure, and reusable template systems with concrete examples. - Ready-to-Use Prompt Library: Role-based prompts (Expert Developer, Code Reviewer, System Architect) and task templates for debugging, refactoring, testing, and security review. - Quality Assurance: Checklists covering anti-patterns (vague instructions, kitchen-sink prompts), prompt injection mitigation, and an iterative optimization workflow. - Use Case: You are building a support ticket classifier. Use the few-shot pattern from the techniques guide with 2-5 diverse input/output examples, then validate the final prompt against the improvement checklist before deploying it. ## Quick Start Ask the AI to design a structured system prompt for your task using the Role-Context-Instructions-Constraints-Examples-Output pattern and validate it against the prompt improvement checklist.
