Agent skill

pdf

Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.

View SKILL.md on GitHub Repository

Stars 3,367

Forks 88

Install this agent skill to your Project

npx add-skill https://github.com/UnicomAI/wanwu/tree/main/configs/microservice/bff-service/configs/agent-skills/wanwu/pdf

SKILL.md

PDF Processing Guide

Overview

This guide covers essential PDF processing operations using Python libraries and command-line tools. For advanced features, JavaScript libraries, and detailed examples, see REFERENCE.md. If you need to fill out a PDF form, read FORMS.md and follow its instructions.

⚠️ Important: Avoid Common Mistakes

Problem: PDF content overflow, missing text, no line wrapping

Root Cause: Using drawString() instead of Paragraph()

❌ Wrong:

python

c = canvas.Canvas("output.pdf")
c.drawString(100, 750, "Long text will overflow and be lost...")

✅ Correct:

python

from reportlab.platypus import SimpleDocTemplate, Paragraph
from reportlab.lib.styles import ParagraphStyle

doc = SimpleDocTemplate("output.pdf")
style = ParagraphStyle('Custom', fontName='STSong-Light', fontSize=11)
story = [Paragraph("Long text will wrap automatically", style)]
doc.build(story)

📖 See: BEST_PRACTICES.md for detailed solutions

Quick Start

Creating PDFs (Recommended Approach)

python

from reportlab.platypus import SimpleDocTemplate, Paragraph
from reportlab.lib.styles import ParagraphStyle
from reportlab.lib.pagesizes import A4
from reportlab.lib.units import cm
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.cidfonts import UnicodeCIDFont

# Register Chinese font
pdfmetrics.registerFont(UnicodeCIDFont('STSong-Light'))

# Create document
doc = SimpleDocTemplate("output.pdf", pagesize=A4,
                       rightMargin=2*cm, leftMargin=2*cm,
                       topMargin=2*cm, bottomMargin=2*cm)

# Create style
style = ParagraphStyle(
    'CustomBody',
    fontName='STSong-Light',
    fontSize=11,
    leading=16,
    firstLineIndent=22
)

# Build content
story = []
story.append(Paragraph("标题文本", ParagraphStyle('Title', fontName='STSong-Light', fontSize=18)))
story.append(Paragraph("正文内容会自动换行，不会超出页面边界。", style))

# Generate PDF
doc.build(story)

Key Points:

✅ Use Paragraph for automatic line wrapping
✅ Use SimpleDocTemplate for proper margins
✅ Register STSong-Light for Chinese text
✅ Use Times-Roman for English text (PDF built-in, no registration needed)

Reading PDFs

python

from pypdf import PdfReader

# Read a PDF
reader = PdfReader("document.pdf")
print(f"Pages: {len(reader.pages)}")

# Extract text
text = ""
for page in reader.pages:
    text += page.extract_text()

Chinese Font Support

This skill includes intelligent font support with automatic font selection for Chinese and English text.

Available Fonts

Primary Fonts (Recommended):

STSong-Light (宋体) - CIDFont, built-in to PDF readers, no file needed
Times-Roman (新罗马) - PDF built-in font, similar to Times New Roman, no file needed

Fallback Fonts (TTF/OTF):

Noto Serif CJK SC (宋体) - OTF version, requires font file
Noto Sans CJK (黑体) - Alternative Chinese font
WenQuanYi Zen Hei (文泉驿正黑) - Alternative Chinese font
WenQuanYi Micro Hei (文泉驿微米黑) - Lightweight Chinese font
Liberation Serif - TTF version of Times New Roman

Why CIDFont is Better

STSong-Light (CIDFont) advantages:

✅ No font files needed - Built into PDF readers
✅ Smaller PDF files - No font embedding required
✅ Perfect compatibility - Standard Adobe CID font
✅ Full Unicode support - Complete Chinese character set
✅ Faster processing - No font file loading
✅ Avoids TTC errors - No TTC format compatibility issues

⚠️ Important: TTC File Format Issue

Problem: ReportLab does not support TTC (TrueType Collection) font files.

Error: TTFError: TTC file "xxx.ttc": postscript outlines are not supported
TTC files like NotoSerifCJK-Regular.ttc cannot be used directly

Solution:

Use CIDFont (STSong-Light) - Recommended, no file needed
Use OTF or TTF format - Supported by ReportLab
This skill automatically skips TTC files - Prevents errors

Supported Font Formats:

Format	Support	Description
CIDFont	✅ Recommended	Built-in to PDF readers
OTF	✅ Supported	OpenType fonts
TTF	✅ Supported	TrueType fonts
TTC	❌ Not supported	TrueType Collection, causes errors

Font Selection Rules

The system automatically selects fonts based on text content:

Chinese characters → STSong-Light (宋体 CIDFont, PDF built-in)
English text and numbers → Times-Roman (新罗马, PDF built-in)

Important: Both fonts are built into PDF readers - no font files needed!

Register Fonts

python

import sys
sys.path.append('scripts')
from register_fonts import register_chinese_fonts, get_chinese_font_name, get_english_font_name

# Register all available fonts
# Priority: CIDFont (STSong-Light) > TTF fonts
registered = register_chinese_fonts()

# Get recommended fonts
chinese_font = get_chinese_font_name()  # STSong-Light (CIDFont, PDF built-in)
english_font = get_english_font_name()  # Times-Roman (PDF built-in)
print(f"Chinese font: {chinese_font}")
print(f"English font: {english_font}")

Note: Both fonts are PDF built-in fonts, no font files needed!

Using Fonts in PDF Creation

With reportlab (Built-in Fonts - Recommended)

python

from reportlab.lib.pagesizes import letter
from reportlab.platypus import SimpleDocTemplate, Paragraph
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.pdfbase import pdfmetrics
from reportlab.pdfbase.cidfonts import UnicodeCIDFont

# Register CIDFont for Chinese (no font file needed!)
pdfmetrics.registerFont(UnicodeCIDFont('STSong-Light'))

# Create document with proper margins
doc = SimpleDocTemplate("output.pdf", pagesize=letter,
                       rightMargin=2*cm, leftMargin=2*cm,
                       topMargin=2*cm, bottomMargin=2*cm)

# Create styles
styles = getSampleStyleSheet()
body_style = ParagraphStyle(
    'CustomBody',
    parent=styles['Normal'],
    fontName='STSong-Light',  # Chinese font
    fontSize=11,
    leading=16,
    firstLineIndent=22  # First line indent
)

# Build content with automatic line wrapping
story = []
story.append(Paragraph("中文内容会自动换行，不会超出页面边界。", body_style))
story.append(Paragraph("English content with Times-Roman will also wrap automatically.", 
                      ParagraphStyle('English', fontName='Times-Roman', fontSize=11)))

doc.build(story)

Key Points:

Use Paragraph instead of drawString for automatic line wrapping
STSong-Light is a CIDFont, requires registration but no font file
Times-Roman is a PDF built-in font, no registration needed
Both fonts are built into PDF readers, no font files needed!

⚠️ Common Mistakes to Avoid

❌ Wrong: Using drawString (no auto-wrap)

python

# This will cause text to overflow page boundaries
c = canvas.Canvas("output.pdf", pagesize=letter)
c.setFont('STSong-Light', 12)
c.drawString(100, 750, "很长的文本不会自动换行，会超出页面边界导致内容丢失...")

✅ Correct: Using Paragraph (auto-wrap)

python

# This will automatically wrap text within margins
doc = SimpleDocTemplate("output.pdf", pagesize=letter)
story = [Paragraph("很长的文本会自动换行，不会超出页面边界。", body_style)]
doc.build(story)

With reportlab (TTF - Fallback)

python

from reportlab.pdfbase.ttfonts import TTFont

# Register TTF font (requires font file)
pdfmetrics.registerFont(TTFont('NotoSerifCJK', '/usr/share/fonts/opentype/noto/NotoSerifCJK-Regular.ttc'))
pdfmetrics.registerFont(TTFont('TimesNewRoman', '/usr/share/fonts/truetype/liberation/LiberationSerif-Regular.ttf'))

# Use in canvas
c = canvas.Canvas("mixed.pdf", pagesize=letter)
c.setFont('NotoSerifCJK', 14)
c.drawString(100, 750, "中文内容")
c.setFont('TimesNewRoman', 14)
c.drawString(100, 720, "English Content 123")
c.save()

With pypdf (Form Filling)

When filling PDF forms, the system automatically detects text type and uses the appropriate font:

python

# The fill_pdf_form_with_annotations.py script automatically:
# 1. Detects Chinese characters in text
# 2. Uses STSong-Light (CIDFont) for Chinese text
# 3. Uses Times-Roman for English text and numbers

# In your fields.json:
{
  "form_fields": [
    {
      "entry_text": {
        "text": "张三",  // Will use STSong-Light (CIDFont)
        "font_size": 12
      }
    },
    {
      "entry_text": {
        "text": "John Smith",  // Will use Times-Roman
        "font_size": 12
      }
    }
  ]
}

Advanced Features

Creating PDFs from Markdown (Optional)

If you have markdown content and want to convert it to PDF with automatic formatting:

python

from markdown_to_pdf import markdown_to_pdf, markdown_file_to_pdf

# From markdown text
markdown_text = """
# 标题

## 二级标题

正文内容会自动换行...

**粗体文本** 和普通文本。
"""

markdown_to_pdf(markdown_text, "output.pdf", title="文档标题")

# Or from markdown file
markdown_file_to_pdf("input.md", "output.pdf")

Features:

✅ Automatic line wrapping
✅ Markdown format support (headers, bold, lists)
✅ Proper page margins
✅ Chinese and English font support

Note: This is an optional helper tool. For most cases, use the direct Paragraph approach shown above.

Common Tasks

To see all available fonts in the system:

bash

# Run from the pdf skill directory
python scripts/register_fonts.py

This will display all registered fonts and the recommended fonts for Chinese and English text.

Python Libraries

pypdf - Basic Operations

Merge PDFs

python

from pypdf import PdfWriter, PdfReader

writer = PdfWriter()
for pdf_file in ["doc1.pdf", "doc2.pdf", "doc3.pdf"]:
    reader = PdfReader(pdf_file)
    for page in reader.pages:
        writer.add_page(page)

with open("merged.pdf", "wb") as output:
    writer.write(output)

Split PDF

python

reader = PdfReader("input.pdf")
for i, page in enumerate(reader.pages):
    writer = PdfWriter()
    writer.add_page(page)
    with open(f"page_{i+1}.pdf", "wb") as output:
        writer.write(output)

Extract Metadata

python

reader = PdfReader("document.pdf")
meta = reader.metadata
print(f"Title: {meta.title}")
print(f"Author: {meta.author}")
print(f"Subject: {meta.subject}")
print(f"Creator: {meta.creator}")

Rotate Pages

python

reader = PdfReader("input.pdf")
writer = PdfWriter()

page = reader.pages[0]
page.rotate(90)  # Rotate 90 degrees clockwise
writer.add_page(page)

with open("rotated.pdf", "wb") as output:
    writer.write(output)

pdfplumber - Text and Table Extraction

Extract Text with Layout

python

import pdfplumber

with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        text = page.extract_text()
        print(text)

Extract Tables

python

with pdfplumber.open("document.pdf") as pdf:
    for i, page in enumerate(pdf.pages):
        tables = page.extract_tables()
        for j, table in enumerate(tables):
            print(f"Table {j+1} on page {i+1}:")
            for row in table:
                print(row)

Advanced Table Extraction

python

import pandas as pd

with pdfplumber.open("document.pdf") as pdf:
    all_tables = []
    for page in pdf.pages:
        tables = page.extract_tables()
        for table in tables:
            if table:  # Check if table is not empty
                df = pd.DataFrame(table[1:], columns=table[0])
                all_tables.append(df)

# Combine all tables
if all_tables:
    combined_df = pd.concat(all_tables, ignore_index=True)
    combined_df.to_excel("extracted_tables.xlsx", index=False)

reportlab - Create PDFs

Basic PDF Creation

python

from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas

c = canvas.Canvas("hello.pdf", pagesize=letter)
width, height = letter

# Add text
c.drawString(100, height - 100, "Hello World!")
c.drawString(100, height - 120, "This is a PDF created with reportlab")

# Add a line
c.line(100, height - 140, 400, height - 140)

# Save
c.save()

Create PDF with Multiple Pages

python

from reportlab.lib.pagesizes import letter
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, PageBreak
from reportlab.lib.styles import getSampleStyleSheet

doc = SimpleDocTemplate("report.pdf", pagesize=letter)
styles = getSampleStyleSheet()
story = []

# Add content
title = Paragraph("Report Title", styles['Title'])
story.append(title)
story.append(Spacer(1, 12))

body = Paragraph("This is the body of the report. " * 20, styles['Normal'])
story.append(body)
story.append(PageBreak())

# Page 2
story.append(Paragraph("Page 2", styles['Heading1']))
story.append(Paragraph("Content for page 2", styles['Normal']))

# Build PDF
doc.build(story)

Subscripts and Superscripts

IMPORTANT: Never use Unicode subscript/superscript characters (₀₁₂₃₄₅₆₇₈₉, ⁰¹²³⁴⁵⁶⁷⁸⁹) in ReportLab PDFs. The built-in fonts do not include these glyphs, causing them to render as solid black boxes.

Instead, use ReportLab's XML markup tags in Paragraph objects:

python

from reportlab.platypus import Paragraph
from reportlab.lib.styles import getSampleStyleSheet

styles = getSampleStyleSheet()

# Subscripts: use <sub> tag
chemical = Paragraph("H<sub>2</sub>O", styles['Normal'])

# Superscripts: use <super> tag
squared = Paragraph("x<super>2</super> + y<super>2</super>", styles['Normal'])

For canvas-drawn text (not Paragraph objects), manually adjust font the size and position rather than using Unicode subscripts/superscripts.

Command-Line Tools

pdftotext (poppler-utils)

bash

# Extract text
pdftotext input.pdf output.txt

# Extract text preserving layout
pdftotext -layout input.pdf output.txt

# Extract specific pages
pdftotext -f 1 -l 5 input.pdf output.txt  # Pages 1-5

qpdf

bash

# Merge PDFs
qpdf --empty --pages file1.pdf file2.pdf -- merged.pdf

# Split pages
qpdf input.pdf --pages . 1-5 -- pages1-5.pdf
qpdf input.pdf --pages . 6-10 -- pages6-10.pdf

# Rotate pages
qpdf input.pdf output.pdf --rotate=+90:1  # Rotate page 1 by 90 degrees

# Remove password
qpdf --password=mypassword --decrypt encrypted.pdf decrypted.pdf

pdftk (if available)

bash

# Merge
pdftk file1.pdf file2.pdf cat output merged.pdf

# Split
pdftk input.pdf burst

# Rotate
pdftk input.pdf rotate 1east output rotated.pdf

Common Tasks

Extract Text from Scanned PDFs

python

# Requires: pip install pytesseract pdf2image
import pytesseract
from pdf2image import convert_from_path

# Convert PDF to images
images = convert_from_path('scanned.pdf')

# OCR each page
text = ""
for i, image in enumerate(images):
    text += f"Page {i+1}:\n"
    text += pytesseract.image_to_string(image)
    text += "\n\n"

print(text)

Add Watermark

python

from pypdf import PdfReader, PdfWriter

# Create watermark (or load existing)
watermark = PdfReader("watermark.pdf").pages[0]

# Apply to all pages
reader = PdfReader("document.pdf")
writer = PdfWriter()

for page in reader.pages:
    page.merge_page(watermark)
    writer.add_page(page)

with open("watermarked.pdf", "wb") as output:
    writer.write(output)

Extract Images

bash

# Using pdfimages (poppler-utils)
pdfimages -j input.pdf output_prefix

# This extracts all images as output_prefix-000.jpg, output_prefix-001.jpg, etc.

Password Protection

python

from pypdf import PdfReader, PdfWriter

reader = PdfReader("input.pdf")
writer = PdfWriter()

for page in reader.pages:
    writer.add_page(page)

# Add password
writer.encrypt("userpassword", "ownerpassword")

with open("encrypted.pdf", "wb") as output:
    writer.write(output)

Quick Reference

Task	Best Tool	Command/Code
Merge PDFs	pypdf	`writer.add_page(page)`
Split PDFs	pypdf	One page per file
Extract text	pdfplumber	`page.extract_text()`
Extract tables	pdfplumber	`page.extract_tables()`
Create PDFs	reportlab	Canvas or Platypus
Command line merge	qpdf	`qpdf --empty --pages ...`
OCR scanned PDFs	pytesseract	Convert to image first
Fill PDF forms	pdf-lib or pypdf (see FORMS.md)	See FORMS.md

Next Steps

For advanced pypdfium2 usage, see REFERENCE.md
For JavaScript libraries (pdf-lib), see REFERENCE.md
If you need to fill out a PDF form, follow the instructions in FORMS.md
For troubleshooting guides, see REFERENCE.md

Maintainer

UnicomAI Core maintainer

Source details

Full Name: UnicomAI/wanwu
Branch: main
Path in repo: configs/microservice/bff-service/configs/agent-skills/wanwu/pdf
License: Apache License 2.0
Topics: agent ai development mcp workflow llm ai-agent agentic-ai golang agentic-framework genai rag ai-agents-framework ai-agent-development-framework open-ai wanwu

Featured Tools

Join Our Newsletter

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).

3,367 88

Explore

Didn't find tool you were looking for?

Search AI Tools

Install this agent skill to your Project

SKILL.md

PDF Processing Guide

Overview

⚠️ Important: Avoid Common Mistakes

Quick Start

Creating PDFs (Recommended Approach)

Reading PDFs

Chinese Font Support

Available Fonts

Why CIDFont is Better

⚠️ Important: TTC File Format Issue

Font Selection Rules

Register Fonts

Using Fonts in PDF Creation

With reportlab (Built-in Fonts - Recommended)

⚠️ Common Mistakes to Avoid

With reportlab (TTF - Fallback)

With pypdf (Form Filling)

Advanced Features

Creating PDFs from Markdown (Optional)

Common Tasks

Python Libraries

pypdf - Basic Operations

Merge PDFs

Split PDF

Extract Metadata

Rotate Pages

pdfplumber - Text and Table Extraction

Extract Text with Layout

Extract Tables

Advanced Table Extraction

reportlab - Create PDFs

Basic PDF Creation

Create PDF with Multiple Pages

Subscripts and Superscripts

Command-Line Tools

pdftotext (poppler-utils)

qpdf

pdftk (if available)

Common Tasks

Extract Text from Scanned PDFs

Add Watermark

Extract Images

Password Protection

Quick Reference

Next Steps

Recommended Agent Skills

agent-stream-nesting-logic

kimi-pdf

kimi-xlsx

doc-coauthoring

internal-comms

mcp-builder