| name | prompt-engineering |
| description | Write effective prompts for LLMs — structure, few-shot examples, chain-of-thought, system prompts, and output parsing. |
| user-invocable | true |
When to use
- Designing or refining prompts for LLM tasks (classification, generation, extraction)
- User asks for better prompts, few-shot examples, or structured output formats
- Building system prompts, guardrails, or output parsers for agent workflows
On-demand loading: Read this skill only when the task clearly matches the description above or trigger phrases below. Do not load for unrelated work.
Prompt Engineering
Write prompts that get reliable, high-quality output from LLMs.
Workflow
- Define the task, desired output format, and failure modes
- Pick techniques from
references/techniques-and-patterns.md (system prompt, few-shot, CoT, structured output)
- Draft a focused prompt with 1–2 examples when format matters
- Test edge cases; iterate — prompts are code
References
| File | Contents |
|---|
references/techniques-and-patterns.md | Core principles, system/few-shot/CoT/structured prompts, code patterns, anti-patterns, tips |
