Claudesmith
Paste a prompt. Get a deep, framework-aware assessment of where it’s strong and where it leaks.
Prompt input
Usage context(optional)
About Claudesmith
Claudesmith is a prompt quality assessment tool for prompt engineers and AI builders. It scores prompts across five dimensions — clarity, specificity, context, constraints, and output format — and recommends the best-fit prompt engineering framework from a library of 27, including CO-STAR, RISEN, RACE, CRISPE, Chain-of-Thought, ReAct, Tree-of-Thought, Self-Refine, Plan-and-Solve, Chain-of-Density, Step-Back, and Skeleton-of-Thought.
How it works
Paste a prompt and run a Quick Assess for instant heuristic scoring with no LLM call, or a Deep Assess for a framework-aware analysis streamed from Claude. Re-assessing a revised prompt produces a delta analysis so you can see whether the revision actually improved things.
Intent categories
Prompts are first classified into one of seven intents — create, transform, reason, critique, recover, clarify, or agentic — which drives framework recommendation.