# AI-Q Prompt Library

Zero-install judgment reviews. Paste a prompt into any capable chat model (Claude,
ChatGPT, Gemini, etc.), then paste your AI-assisted work. The model runs the AI-Q protocol
— work map → interview → AI-Q — by hand.

These are the human-readable counterpart to the [SDK](../sdk/typescript/): same rubric,
same six Discernment dimensions, same scoring. Use them to understand the framework, demo
it, or run a review without writing code. (The prompts score Discernment — the spine of
[AI-Q](../spec/aiq.md); a six-dimension review's AI-Q equals its Discernment sub-score.)

| Prompt | Domain |
|---|---|
| [coding-review.md](coding-review.md) | Code |
| [writing-review.md](writing-review.md) | Writing & communication |
| [research-review.md](research-review.md) | Research & analysis |
| [universal-review.md](universal-review.md) | Any work type |

**The one rule:** answer the interview questions yourself, *before* asking the AI to
answer them for you. The moment the AI answers them, you've recreated the exact problem
Hone measures.

> Other domains (product, strategy) use the same structure with the matching
> [rubric](../rubrics/). Contributions of domain prompts welcome — see
> [CONTRIBUTING](../CONTRIBUTING.md).