Run2 pdf skill form filling

Filling PDF Forms with the PDF Skill

Overview

The PDF skill provides specialized Python scripts for filling PDF forms that have fillable form fields. This is the recommended approach for working with AcroForm PDFs.

Prerequisite Check

Before attempting to fill a form, verify the PDF has fillable fields:

python3 scripts/check_fillable_fields.py <pdf_file>

• If output is "This PDF has fillable form fields" → proceed with steps below
• If output is "This PDF does not have fillable form fields" → use annotation-based approach instead

Workflow for Fillable Forms

Step 1: Extract Form Field Information

Extract all form fields and their metadata:

python3 scripts/extract_form_field_info.py <input.pdf> <output_field_info.json>

This creates a JSON file with all fields including:

• field_id: Unique identifier for the field (use this in field_values.json)
• type: "text", "checkbox", "radio_group", or "choice"
• page: Page number (1-based)
• rect: Bounding box [left, bottom, right, top] in PDF coordinates
• For checkboxes: checked_value and unchecked_value
• For radio groups/choices: radio_options or choice_options with values

Step 2: Create field_values.json

Create a JSON file with values for each field to fill:

[
  {
    "field_id": "SC-100[0].Page2[0].List1[0].Item1[0].PlaintiffName1[0]",
    "description": "Plaintiff's name",
    "page": 2,
    "value": "John Smith"
  },
  {
    "field_id": "SomeCheckboxField",
    "description": "Agreement checkbox",
    "page": 1,
    "value": "/1"  // Use the "checked_value" from field_info for checked checkboxes
  },
  {
    "field_id": "RadioGroupField",
    "description": "Choose option",
    "page": 1,
    "value": "/Yes"  // Use one of the "value" entries from radio_options
  }
]

Key Rules:

• field_id: Must match exactly from field_info.json (case-sensitive)
• value: For text fields, use a string; for checkboxes/radio, use the checked_value or option value
• page: Must match the page number in field_info.json
• description: Optional, but helpful for documentation

Step 3: Fill the Form

Run the fill script to create the filled PDF:

python3 scripts/fill_fillable_fields.py <input.pdf> <field_values.json> <output.pdf>

The script will:

• Verify all field_ids and values are valid
• Print error messages if there are issues
• Create the output PDF if successful

Important Notes

Field ID Matching

• Field IDs are case-sensitive and must be exact matches
• Field IDs use XPath-like notation (e.g., SC-100[0].Page2[0].List1[0].Item1[0].PlaintiffName1[0])
• Do not modify or abbreviate field IDs

Value Formats

• Text fields: Plain strings (no special formatting needed)
• Checkboxes: Use the checked_value when you want to check; omit the field or use unchecked_value to leave unchecked
• Radio buttons: Use one of the values from the radio_options list
• Dropdowns: Use one of the values from the choice_options list

Common Issues

• If script reports "Invalid field_id" → copy the exact ID from field_info.json
• If checkbox doesn't appear checked → use the correct checked_value (often "/1" or "/2")
• If script reports "Invalid value for field" → verify the value matches an option from field_info

Optional Fields

• To leave a field empty, simply omit it from field_values.json
• Court-filled fields (like case number, trial date) should be left empty/omitted
• Don't include signature or date fields that the court fills in

File Path Usage

The scripts must be run from the /root/.claude/skills/pdf directory or with full paths:

cd /root/.claude/skills/pdf
python3 scripts/fill_fillable_fields.py /path/to/input.pdf /path/to/field_values.json /path/to/output.pdf