Introduction
Prompt Architect is a Claude Code skill that systematically transforms vague prompts into structured, expert-grade prompts using 27 research-backed frameworks. It scores every prompt across 5 dimensions (clarity, specificity, context, completeness, structure) and recommends the best framework for your use case. Supports 30+ tools including Claude Code, ChatGPT, Cursor, and Gemini CLI. GitHub 96+ stars, MIT open source.
27 Research-Driven Prompt Frameworks
27 frameworks grouped into three tiers by complexity:
High complexity: CO-STAR (content creation), RISEN (multi-step processes), CRISPE (creative writing), BROKE (constrained tasks)
Medium complexity: 17 frameworks including TIDD-EC (precise tasks), Chain of Thought (step-by-step reasoning), Tree of Thought (branch exploration), ReAct (reasoning + action)
Low complexity: 7 frameworks including RTF (quick tasks), BAB (transformation tasks), APE (automatic prompt engineering)
Workflow — 5-Stage Pipeline
- Analyze — score the prompt across 5 dimensions, 1–10 each
- Recommend — propose the best framework with reasoning
- Dialogue — ask 3–5 clarifying questions
- Apply — rewrite with the chosen framework
- Iterate — refine until satisfied