Agent Skills: PDF Skill

PDF Skill

When to use

• Read or review PDF content where layout and visuals matter.
• Create PDFs programmatically with reliable formatting.
• Validate final rendering before delivery.

Workflow

1. Prefer visual review: render PDF pages to PNGs and inspect them.
   • Use pdftoppm if available.
   • If unavailable, install Poppler or ask the user to review the output locally.
2. Use reportlab to generate PDFs when creating new documents.
3. Use pdfplumber (or pypdf) for text extraction and quick checks; do not rely on it for layout fidelity.
4. After each meaningful update, re-render pages and verify alignment, spacing, and legibility.

Temp and output conventions

• Use tmp/pdfs/ for intermediate files; delete when done.
• Write final artifacts under output/pdf/ when working in this repo.
• Keep filenames stable and descriptive.

Dependencies (install if missing)

Prefer uv for dependency management.

Python packages:

uv pip install reportlab pdfplumber pypdf

If uv is unavailable:

python3 -m pip install reportlab pdfplumber pypdf

System tools (for rendering):

# macOS (Homebrew)
brew install poppler

# Ubuntu/Debian
sudo apt-get install -y poppler-utils

If installation isn't possible in this environment, tell the user which dependency is missing and how to install it locally.

Environment

No required environment variables.

Rendering command

pdftoppm -png $INPUT_PDF $OUTPUT_PREFIX

Quality expectations

• Maintain polished visual design: consistent typography, spacing, margins, and section hierarchy.
• Avoid rendering issues: clipped text, overlapping elements, broken tables, black squares, or unreadable glyphs.
• Charts, tables, and images must be sharp, aligned, and clearly labeled.
• Use ASCII hyphens only. Avoid U+2011 (non-breaking hyphen) and other Unicode dashes.
• Citations and references must be human-readable; never leave tool tokens or placeholder strings.

Final checks

• Do not deliver until the latest PNG inspection shows zero visual or formatting defects.
• Confirm headers/footers, page numbering, and section transitions look polished.
• Keep intermediate files organized or remove them after final approval.

Filling non-fillable PDFs

Use PyMuPDF (fitz) to insert text and images at precise positions on PDFs that lack interactive form fields.

Additional dependency

uv pip install pymupdf

Workflow

1. Analyze structure — Extract text positions to map form layout:

import fitz
doc = fitz.open(pdf_path)
for page in doc:
    text_dict = page.get_text("dict")
    for block in text_dict["blocks"]:
        if "lines" in block:
            for line in block["lines"]:
                for span in line["spans"]:
                    text = span["text"].strip()
                    if text:
                        bbox = span["bbox"]  # (x0, y0, x1, y1)
                        print(f"'{text}' at y={bbox[1]:.1f} x={bbox[0]:.1f}")

2. Calculate fill positions — Form boxes are typically 18-25pt below their labels. PDF coordinates: origin top-left, 72 pts/inch, Y increases downward. US Letter = 612x792pt.

3. Insert text:

page.insert_text((x, y), "value", fontsize=11, color=(0, 0, 0))

4. Insert images (signatures):

page.insert_image(fitz.Rect(x0, y0, x1, y1), filename="signature.png")

Use PNG with transparent background. Maintain aspect ratio.

Troubleshooting

| Issue | Fix | |-------|-----| | Text in wrong position | Re-run analysis, add 18-25pt to label y for input box | | Text too large/small | Default 11pt; use 7-8pt for dense tables | | Signature doesn't fit | Check aspect ratio matches rect proportions | | Text overlaps content | Reduce fontsize, verify inserting in box not on label |

Why PyMuPDF over other libraries

PyPDF2/fillpdf only work with fillable forms. reportlab creates new PDFs (can't overlay). PyMuPDF inserts text/images at any position on existing PDFs.

Signing and filling documents

Complete workflow for filling forms and adding e-signatures to PDF and Word documents.

Scripts location: ~/.claude/scripts/sign-document/ Signature assets: ~/.claude/scripts/sign-document/assets/

Additional dependencies

pip install pypdf reportlab pdf2image python-docx cairosvg pillow
brew install poppler  # Required for pdf2image on macOS

Decision tree

Input document?
├── PDF
│   ├── Has fillable fields? → Fillable PDF Workflow
│   └── No fillable fields? → Annotation-Based Workflow
└── Word (.docx) → Word Document Workflow

All workflows end with → Add Signature → Export as PDF

Step 1: Detect document type

file "$DOCUMENT_PATH"

Step 2: Check for fillable fields (PDF only)

python ~/.claude/scripts/sign-document/check_fillable_fields.py "$INPUT_PDF"

Fillable PDF workflow

# Extract field info
python ~/.claude/scripts/sign-document/extract_form_field_info.py "$INPUT_PDF" field_info.json

# Convert to images for visual reference
python ~/.claude/scripts/sign-document/convert_pdf_to_images.py "$INPUT_PDF" form_images/

# Fill the form (create field_values.json first)
python ~/.claude/scripts/sign-document/fill_fillable_fields.py "$INPUT_PDF" field_values.json filled.pdf

field_values.json format:

[
  {"field_id": "last_name", "description": "Last name", "page": 1, "value": "Cai"},
  {"field_id": "agree_checkbox", "description": "Agreement", "page": 2, "value": "/On"}
]

Value formats: Text → string, Checkbox → "/On" or "/Yes", Radio → value from radio_options, Dropdown → value from choice_options.

Annotation-based workflow (non-fillable PDFs)

# Convert to images, analyze layout, create fields.json
python ~/.claude/scripts/sign-document/convert_pdf_to_images.py "$INPUT_PDF" form_images/

# Validate bounding boxes
python ~/.claude/scripts/sign-document/check_bounding_boxes.py fields.json
python ~/.claude/scripts/sign-document/create_validation_image.py 1 fields.json form_images/page_1.png validation.png

# Fill
python ~/.claude/scripts/sign-document/fill_pdf_form_with_annotations.py "$INPUT_PDF" fields.json filled.pdf

Bounding box format: [left, top, right, bottom] in pixels. Label and entry boxes MUST NOT overlap.

Word document workflow

from docx import Document
doc = Document('template.docx')
replacements = {'{{name}}': 'Jeremy Cai', '{{date}}': 'January 13, 2026'}
for paragraph in doc.paragraphs:
    for key, value in replacements.items():
        if key in paragraph.text:
            for run in paragraph.runs:
                if key in run.text:
                    run.text = run.text.replace(key, value)
doc.save('filled.docx')

Add signature and export (final step)

python ~/.claude/scripts/sign-document/add_signature_and_export.py filled.pdf output.pdf

| Option | Default | Description | |--------|---------|-------------| | --position | bottom-right | bottom-left, bottom-right, bottom-center | | --margin | 50 | Points from page edge | | --width | 200 | Signature width in points | | --page | -1 (last) | Page to sign (0-indexed, -1=last) | | --no-signature | false | Convert to PDF without signature |

User data reference

When filling documents for Jeremy Cai:

• Full Name: Jeremy Cai
• DOB: April 21, 1995
• Height: 182cm / 6'0"
• Locations: Park City UT, Chicago IL, Los Angeles CA
• Occupation: Serial Entrepreneur & Angel Investor
• Companies: Fountain, Italic, Carbon (acquired by Perplexity), Blemish, Courtly
• Spouse: Katherine (Kati) Holland Cai (b. Mar 23, 1994)
• Parents: Emily Chen (b. Apr 16, 1963), James Cai (b. Nov 10, 1963)
• Sister: Jillian Cai (b. Mar 7, 1998)