Enhance Prompt

What Is This?

Overview

Enhance Prompt is a skill designed for the Google Labs Stitch platform that transforms rough, vague UI generation ideas into polished, structured prompts optimized for better output. When working with AI-powered UI generation tools, the quality of your input directly determines the quality of your output. A poorly worded prompt produces generic, misaligned results, while a well-structured prompt produces precise, usable UI components.

This skill acts as a prompt engineering layer between your initial idea and the Stitch generation engine. It enhances specificity, injects relevant UI/UX terminology, adds design system context, and restructures your raw input into a format that Stitch can interpret with greater accuracy. The result is a refined prompt that communicates layout intent, component hierarchy, visual style, and interaction patterns clearly.

The skill uses Read and Write tools to process your input and produce an enhanced output file. It is particularly useful for teams and individuals who have strong design intuitions but limited experience writing structured AI prompts, bridging the gap between creative intent and technical execution.

Who Should Use This

• Product designers who want to rapidly prototype UI screens without writing detailed technical prompts from scratch
• Frontend developers who need to generate UI mockups quickly and want consistent, component-aware output from Stitch
• UX researchers who want to visualize interface concepts for usability testing without deep design tool expertise
• Product managers who need to communicate UI ideas to engineering teams using generated visual references

Why Use It?

Problems It Solves

• Vague prompts produce unusable output. Generic descriptions like "make a dashboard" give Stitch insufficient context, resulting in layouts that do not match your actual product needs.
• Missing UI/UX vocabulary limits generation accuracy. Without terms like "card grid," "sticky navigation," "modal overlay," or "skeleton loader," the generation engine cannot infer the correct component types.
• Design system context is lost. Raw prompts rarely include spacing scales, typography hierarchies, or color token references, causing generated UI to drift from your established system.
• Inconsistent prompt structure wastes iteration cycles. Without a reliable structure, each generation attempt requires manual trial and error to refine.

Core Highlights

• Automatically enhances vague input with specific UI/UX terminology
• Injects design system context including layout patterns and component types
• Structures prompts for optimal Stitch generation performance
• Reduces iteration cycles by improving first-pass output quality
• Supports both simple component prompts and complex multi-section layouts
• Works with Read and Write tools for seamless file-based workflows
• Bridges the gap between non-technical intent and technical prompt requirements
• Produces consistent, repeatable prompt formats across projects

How to Use It?

Basic Usage

Provide a rough UI description as your input. The skill reads your input, processes it through prompt engineering logic, and writes an enhanced version.

Input: "a settings page for a mobile app"

Enhanced Output:
"Design a mobile settings screen with a vertically scrolled list layout.
Include grouped sections: Account, Notifications, Privacy, and Support.
Each section uses a card container with 16px padding, divider lines between
rows, and right-aligned chevron icons for navigation. Use a sans-serif
typeface at 14px body size. Apply a light background (#F5F5F5) with white
card surfaces. Include a sticky header with back navigation and page title."

Specific Scenarios

Scenario 1: Component-level prompt enhancement A developer needs a button component prompt. Raw input: "primary button." Enhanced output specifies size variants, states (default, hover, disabled, loading), border radius, padding values, and label typography.

Scenario 2: Full-page layout prompt A product manager describes "a checkout page." The skill expands this into a structured prompt covering form sections, order summary panel, progress indicator, CTA placement, and error state handling.

Real-World Examples

• Transforming "user profile page" into a detailed prompt specifying avatar placement, bio section, activity feed, and follow button hierarchy
• Converting "admin table view" into a prompt with sortable columns, pagination controls, row actions, and empty state messaging

When to Use It?

Use Cases

• Rapid UI prototyping at the start of a product sprint
• Generating reference screens for stakeholder presentations
• Creating component variations for design system documentation
• Producing UI mockups for usability testing sessions
• Translating written product requirements into visual generation prompts
• Standardizing prompt quality across a multi-person design team
• Accelerating handoff by generating developer-ready component specifications

Important Notes

Requirements

• Access to the Google Labs Stitch platform is required to use the generated prompts
• The skill requires Read and Write tool permissions to process input and output files
• A basic understanding of your target design system improves enhancement accuracy
• Input descriptions should be in English for optimal processing results