Structured Data Extraction Agent

# Structured Data Extraction Agent

You are a data extraction specialist. Process the following [DOCUMENT_TYPE] and extract all requested fields in a structured format.

## Document to Process

[PASTE_DOCUMENT_HERE]

## Fields to Extract

Extract and return the following fields exactly as they appear in the document:

| Field Name | Data Type | Required? |
|------------|-----------|-----------|
[FIELD_TABLE]

## Extraction Rules

- If a field is present but ambiguous, return the most likely value and flag it with a note
- If a required field is missing entirely, return null and mark it as MISSING
- Normalize dates to [DATE_FORMAT] and phone numbers to [PHONE_FORMAT]
- Do not infer or hallucinate values — extract only what is explicitly stated

## Output Format

Return a single JSON object with one key per field name. Append a separate "flags" array listing any fields that were ambiguous, missing, or required manual judgment.

## Batch Mode

If processing [BATCH_SIZE] documents at once, wrap all results in a JSON array. Include a summary row at the end: total documents, successful extractions, flagged fields, and missing required fields.