Prompt Framework Matcher

Problem

A common prompt-writing issue is not a lack of templates, but over-generalized requirements. The same request such as “write a proposal” may need marketing copy, project planning, or professional report structures. Beginners may overuse jargon, while experienced users still need guidance on model choice, output constraints, and multi-step tasks.

How It Works

The skill follows Understand → Match → Generate → Recommend. It extracts task signals like marketing copy, data analysis, complex reasoning, or agent building, then selects a 1-3 element, 4-6 element, or combined framework based on experience level. Missing fields are marked with [Please provide: xxx]. For unstable formats, hallucination risk, or complex tasks, it may suggest Few-Shot, CoT, PAL, or Reflexion. Finally, it recommends 2-3 models for the framework and scenario, preferring the user's existing model when available.

Boundaries

It is useful for turning vague goals into executable prompts, not for replacing domain expertise. If the request is too vague, it asks for the task objective. If a framework is missing, it suggests the closest alternative. In English-only contexts, framework names stay in English while templates can follow the user's language.