Treat prompts like products: versioned, tested, and measurable.
Quick Start
- Define the job: inputs, outputs, and the “definition of done”.
- Write the smallest prompt that:
- states constraints clearly
- defines output format
- includes edge-case handling
- Add examples only when needed (few-shot is expensive).
- Create an eval set: representative cases + adversarial cases.
- Iterate with diffs: change one thing, measure impact.
Optional tool: scaffold a prompt + eval harness
python ~/.codex/skills/senior-prompt-engineer/scripts/scaffold_prompt_eval.py . --out evals/prompt_eval
References
- Prompt review checklist:
references/prompt-review.md