Prompt Library
Writing your own prompts is the goal — this Library is the toolbox to get you there. Study, adapt, then outgrow it.
- 🔧 Starter — copy/fill to see the pieces.
- 🧰 Skill Builders — structure, tone, constraints.
- ⚙️ Advanced Operators — strict formats, evaluators, tools.
Pick a level to load its prompts. Switch anytime.
How to use the Library
Pick a level, open a prompt, fill the blanks, and send it to Practice. Improve, then reuse.
- Choose a level. Starter → Intermediate → Advanced as you ramp up.
- Open a prompt. Click any card to expand. Scan the intent + variables.
- Fill variables. Replace
{placeholders}with your context (audience, goal, constraints). - Copy or “Open in Practice.” Send it to the Practice page to score + iterate.
- Tighten constraints. Add format, limits, edge cases, evaluation criteria.
- Save your best. Keep winning prompts in your own docs for reuse.
Tip • Start broad → constrain
Tip • Add “must include / must avoid”
Tip • Define success criteria
- Format beats fluff: ask for JSON/table/bullets and set token/word limits.
- Test edge cases: include low-resource, noisy input, or tricky constraints.
- Compare rewrites: in Practice, iterate and keep the version that scores + performs best.
- Promote winners: turn strong prompts into templates for your workflows.
Why not just ask ChatGPT to rewrite?
Rewrites give you one-off text with little rationale. Practice shows what to change and why, so you build prompt craft you can reuse across tasks and models.
