Prompt Engineering

This is the skill you want when you're writing system prompts, building RAG pipelines, or debugging why your LLM keeps hallucinating. It covers chain-of-thought reasoning, few-shot example design, structured output formatting, and the decision tree between prompting versus fine-tuning. The guidance is provider-agnostic but practical, with concrete templates for system prompts and anti-patterns to avoid. Treats prompts like code: versioned, tested against eval sets, and iterated on. If you're chaining multiple LLM calls or trying to coax reliable JSON out of a model, this gives you the patterns and token economics to do it right.

Install to Claude Code

npx -y skills add absolutelyskilled/absolutelyskilled --skill prompt-engineering --agent claude-code