Prompt Engineer

The prompt-engineer skill transforms raw user queries into optimized, context-aware instructions tailored for Claude-based AI agents. It automatically refines vague requests into structured prompts that enhance reasoning accuracy and output quality.

When to use it

• You have a complex coding task but lack the specific technical phrasing required to get the best result from an LLM.
• Your initial prompt generates generic or hallucinated code, and you need to inject logical constraints or step-by-step reasoning.
• You are working with Claude specifically and want to leverage its advanced instruction-following capabilities without writing complex system messages manually.

Key capabilities

• Converts casual language into precise technical directives suitable for code generation.
• Injects implicit context about best practices, error handling, and edge cases based on the request topic.
• Structures output requirements to ensure the AI agent adheres to specific formatting or logic constraints.

Example prompts

• "I need a Python script to scrape data from this site, but make sure it handles rate limiting gracefully."
• "Refine my request for a React component so the AI includes proper TypeScript typing and accessibility attributes."
• "Turn this vague idea for a database migration into a detailed step-by-step plan for an AI agent to execute safely."

Tips & gotchas

This skill is optimized specifically for Claude models; results may vary with other LLMs. For best results, provide clear context about your current environment or constraints before asking the agent to refine your prompt.