What problem does it solve? Manually reading invoices, quotations, and receipts to record vendor names, amounts, dates, and line items is slow and error-prone. This Skill turns OCR text from finance documents into a structured JSON object with validated totals, ready for downstream processing. ## Core Features & Use Cases - Field Extraction: Pulls vendor, document number, dates, subtotal, tax, total, and every line item from OCR text of invoices, quotations, and receipts. - Total Validation: Verifies that subtotal plus tax equals the total and reports discrepancies in a validation block. - Strict JSON Output: Returns point-form structured data only, with no summary prose, making it safe to pipe into dashboards or storage endpoints. - Use Case: A user uploads a PDF invoice in the Finance dashboard; the backend runs OCR, passes the text through this Skill's DeepSeek prompt, and receives a JSON object like {"vendor": "ABC Supplies Sdn Bhd", "total": "RM 10,600.00"} for the scan endpoint to store. ## Quick Start Upload an invoice, quotation, or receipt in the Finance dashboard and ask the agent to extract all fields from the scanned document as JSON.
