Contract Cross-Reference Validation App Builder

## Role

You are a legal technology specialist building contract validation tools. You combine legal document precision with systematic software architecture thinking.

## Task

Build a production-ready React application that validates cross-references in legal contracts. The system must catch section references pointing to non-existent clauses, definitions referencing wrong paragraphs, exhibits cited incorrectly, and other structural errors.

## Core Features

**1. Document Processing**

Parse uploaded contracts (PDF, DOCX, TXT) and extract all cross-references: section references, defined terms, exhibit references, schedule references, and clause citations.

**2. Structure Mapping**

Build a complete document map showing actual structure with all sections, subsections, definitions, exhibits, and schedules with real numbering.

**3. Validation Engine**

Cross-validate every reference against actual document structure using regex patterns that detect sections, defined terms in quotes followed by "means" or "shall mean", exhibit references with letter and optional number suffixes, and nested references like "as defined in Section 3.2(a)(ii)".

**4. Error Classification**

Flag discrepancies by severity based on {{error-threshold}}:
- CRITICAL: non-existent references
- HIGH: incorrect numbering
- MEDIUM: ambiguous references
- LOW: formatting inconsistencies

**5. Interactive Interface**

Three-panel layout:
- Document upload/preview pane
- Live validation results with side-by-side comparison showing extracted reference vs actual target with context snippets
- Document structure tree
- Color-coded reference badges: green (valid), red (broken), yellow (ambiguous)

**6. Professional Reporting**

Generate PDF reports with executive summary, detailed findings table, and remediation recommendations, branded with {{report-branding}}.

## Context

**Tech Stack:** React 18, TypeScript, Tailwind CSS, shadcn/ui, Zustand, react-pdf, mammoth, jsPDF

**Visual Design:** Linear.app-inspired minimalism with dark mode (slate-900 background), emerald-500 success indicators, rose-500 error highlights

**Performance:** Process 100-page contract in under 10 seconds using web workers to avoid UI blocking

**Security:** All processing client-side only, never upload to server, clear data after session

**UX Details:**
- Keyboard shortcuts: Cmd+U (upload), Cmd+R (validate), Cmd+E (export)
- Animated progress indicators showing extraction/mapping/validation stages
- Auto-save validation state every 30 seconds
- User-friendly error messages for upload failures, missing structure, zero references detected

**Legal Accuracy:** Use Bluebook citation standards, recognize legal document patterns, handle peculiar numbering schemes (e.g., Schedule A-1, Section 10.03(b)(iv))

**User Input:** The contract file to validate is {{contract-file}}.

## Output

Provide a single, complete React application organized as:

**Component Hierarchy:**
App wraps ContractValidator, which composes DocumentUploader, ValidationEngine, ResultsPanel, StructureTree, and ReportGenerator.

**Custom Hooks:**
Implement useContractParser for document ingestion, useReferenceExtractor for citation detection, and useValidationEngine for cross-reference checking.

**Code Requirements:**
- Full TypeScript interfaces for ContractDocument, Reference, ValidationError, DocumentStructure, and ValidationReport
- Tailwind utility classes only (no custom CSS files)
- Self-contained code using only the specified libraries
- Include 2-3 sample contracts with intentional errors (broken section references, missing exhibits, mismatched definitions) for demonstration
- Conversational interface language ("Upload Contract" not "Ingest Document", "Found 3 issues" not "3 validation failures detected")
- Gentle animations (fade-in results, smooth panel transitions) and helpful empty states ("Drop a contract here to get started")

Generate the complete application code ready to run, including all component files, hooks, type definitions, utility functions for regex matching and PDF generation, and sample data.