prompt-engineering

name	prompt-engineering
description	Use when designing or improving prompts, commands, hooks, skills, or sub-agent instructions, including output quality, reliability, and token efficiency.

Prompt Engineering

Use this skill to design robust prompts that are clear, testable, and cost-aware.

When to Use This Skill

• Creating or refactoring prompts for skills, commands, hooks, and reusable templates
• Improving output consistency, formatting, or reasoning quality
• Reducing token usage or latency while preserving quality
• Choosing the right level of strictness (high, medium, low freedom)
• Defining guardrails and verification steps for fragile workflows

Workflow

1. Define the exact task outcome, constraints, and required format.
2. Choose the minimum instruction detail needed for the task's fragility.
3. Start with a concise baseline prompt.
4. Add structure: examples, explicit output schema, and failure handling.
5. Add verification criteria for correctness and uncertainty handling.
6. Test on varied and edge-case inputs, then iterate.

Expected Outputs

• A production-ready prompt (or template) with clear task, context, and output contract
• A short rationale for major design choices
• A test checklist with representative inputs and edge cases
• Notes on token and latency tradeoffs

References

• Core Patterns
• Agent Best Practices
• Persuasion Principles