The "CreditGhost" Master Prompt
Role: You are an expert Fintech UI/UX Developer. Project Name: CreditGhost Tagline: Behavioral Credit Scoring for Thin-File Borrowers.
Objective: Build a professional, high-trust React web application with a linear 3-step flow (Input → Loading → Results).
Step 1: Onboarding & Upload (The "Handshake")
Create a centered, clean card with a deep navy/slate theme (#1A2B3C).
Inputs: Name (Text), PAN Number (Masked input with regex validation for 10 alphanumeric chars).
Feature: A Drag-and-Drop PDF Upload zone using a dashed border.
Requirement: The "Analyze Credit Profile" button should only be active once all fields are filled and a file is attached.
Step 2: The "Intelligence Engine" (Loading State)
When "Analyze" is clicked, transition to a full-screen processing state.
Visual: A large, pulsing circular progress ring.
Dynamic Status Text: Cycle through these messages every 800ms:
"Parsing Bank Statement with pdfplumber..."
"Extracting 14 Behavioral Risk Signals..."
"Analyzing Inflow Variance & Essential Spend..."
"Running XGBoost Inference Model..."
"Detecting Income Manipulation Patterns..."
Step 3: The "Risk Intelligence" Dashboard (Results)
Show a professional dashboard using the following JSON-mapped attributes:
Credit Score Gauge: A semi-circle speedometer (0-920). Use Green for 750+, Amber for 600-750, Red for <600.
Confidence/Fraud Shield: A secondary circular meter (0-100%) labeled "Verification Confidence."
Reasoning Feed: Use a vertical list of cards. Use Lucide-React icons (Green Check for positive logic, Red Alert for risks).
Attributes to Display: Income Stability, Savings Buffer Days, and Essential Spending Ratio.
Styling & Vibe:
Theme: "Modern Enterprise Fintech." Use Dark Mode by default.
Colors: Deep Navy (#0F172A), Emerald Green (#10B981), and Amber (#F59E0B).
Typography: Use a clean Sans-Serif (Inter) for UI and Monospace (JetBrains Mono) for the loading logs to give it a "Tech/AI" feel.
Responsiveness: Must look flawless on mobile and desktop.
Pro-Tip for Vibe Coding:
If the AI tool asks for the Backend API, tell it:
"For now, mock the API response with a 3-second delay and return a random Credit Score between 300-920, a Confidence Score between 50-100, and 3 bullet points of reasoning based on the score value."