Prompt Engineering

Prompt Engineering is a standalone skill that teaches any OpenClaw agent how to write prompts that produce consistent, high-quality output from LLMs — not vague instructions that get vague results, but structured prompts with clear roles, specific constraints, and output formats that consistently deliver what you need. Most agents send unstructured requests and hope for the best. This skill replaces hope with methodology.

Drop this skill into any OpenClaw persona's skills folder and your agent immediately knows how to structure prompts for reliable output, use few-shot examples to communicate style, apply chain-of-thought for complex reasoning, specify output formats that the model follows consistently, and optimize prompts for accuracy, consistency, and token efficiency.

✏️ Everything Included This skill covers the full prompt engineering toolkit — no separate purchases needed:

Prompt Structure — the Role + Task + Context + Constraints + Format pattern that works for any prompt Few-Shot Examples — showing the model what you want instead of describing it Chain of Thought — step-by-step reasoning for complex analysis and decision-making Output Formatting — JSON schemas, tables, templates, and structured output that the model follows Domain Patterns — specific prompt structures for code generation, analysis, content creation, and data extraction

Proven patterns:

Role + Task + Context + Constraints + Format structure for any prompt type Few-shot examples that communicate tone and style more effectively than adjectives Negative constraints that tell the model what to avoid, not just what to do Iterative refinement for building complex output in stages

What makes this different from prompt guides:

Agent-native — designed to be read and applied by an AI agent, not a human Actionable patterns — not theory about how LLMs work, but copy-paste prompt structures that produce results Domain-specific templates — separate patterns for code, analysis, content, and data extraction Optimization techniques — reducing token usage, improving consistency, and increasing accuracy Anti-patterns included — the 6 most common prompting mistakes that produce generic or unreliable output

Includes SKILL.md with complete prompt engineering documentation covering core principles, 6 prompt patterns, domain-specific templates, optimization techniques, and anti-patterns. Designed for OpenClaw. Works with any persona. Makes every other skill more effective.