What problem does it solve? AI features embedded in applications often produce inconsistent outputs, hallucinated data, or unreliable JSON because their prompts lack structure, examples, and explicit constraints. This Skill provides a systematic protocol for writing, versioning, and testing prompts so LLM features behave predictably in production. ## Core Features & Use Cases - Structured System Prompts: Builds prompts from four components — role, goal, constraints, and output format — with TypeScript templates for extraction, classification, summarization, RAG, and code generation tasks. - Few-Shot & Chain-of-Thought Patterns: Supplies few-shot example design for extraction and classification, plus structured reasoning formats for multi-step advisory tasks. - RAG & Tool Description Design: Provides context-injection templates with boundary instructions to prevent hallucination, and precise tool description patterns that disambiguate similar functions. - Versioning & Evaluation: Includes a versioned prompt registry with changelogs and an eval harness pattern that tests prompts against fixed input sets. - Use Case: Your invoice extraction feature returns inconsistent dates and invents missing fields. Use this Skill to rewrite the prompt with explicit null handling, ISO date constraints, few-shot edge cases, and an eval suite that measures accuracy before deployment. ## Quick Start Ask the AI to audit your existing system prompt for an extraction or classification feature and rewrite it using the role, goal, constraints, and format structure with few-shot examples.
