AGENT SKILL

pdf-text-extraction

Extract text from PDFs using shell tools when read_file fails

使用场景文档处理
License 许可
MIT
所属仓库
HKUDS/AgentPilot
最近更新
2026-11-12
Stars
★ 8.0k
Forks
944

01 前言 Front Matter

name
pdf-text-extraction
description
Extract text from PDFs using shell tools when read_file fails

02 资源文件 1

03 SKILL.md 正文

PDF Text Extraction (Fallback Method)

When to Use This Skill

Use this skill when read_file with filetype='pdf':

  • Returns binary image data instead of text
  • Produces errors or incomplete content
  • Fails to extract structured data reliably

The built-in PDF handler is unreliable for structured text extraction. Shell-based tools provide more robust alternatives.

Available Methods

Method 1: pdftotext (Recommended)

# Extract text from PDF to stdout
pdftotext /path/to/file.pdf -

# Extract text to a file
pdftotext /path/to/file.pdf output.txt

# Preserve layout (maintains spacing/structure)
pdftotext -layout /path/to/file.pdf output.txt

Usage in agent:

run_shell command="pdftotext -layout /path/to/document.pdf -"

Method 2: pdfinfo (Metadata)

# Get PDF metadata (pages, author, creation date, etc.)
pdfinfo /path/to/file.pdf

Usage in agent:

run_shell command="pdfinfo /path/to/document.pdf"

Method 3: Python with PyMuPDF (fitz)

import fitz  # PyMuPDF

doc = fitz.open("/path/to/file.pdf")
text = ""
for page in doc:
    text += page.get_text()
doc.close()
print(text)

Usage in agent:

run_shell command="python3 -c \"import fitz; doc=fitz.open('file.pdf'); print(''.join(p.get_text() for p in doc))\""

Method 4: Python with pdfplumber (Tables)

import pdfplumber

with pdfplumber.open("/path/to/file.pdf") as pdf:
    for page in pdf.pages:
        text = page.extract_text()
        tables = page.extract_tables()  # For tabular data

Usage in agent:

run_shell command="python3 -c \"import pdfplumber; pdf=pdfplumber.open('file.pdf'); print(''.join(p.extract_text() or '' for p in pdf.pages))\""

Workflow

  1. Try pdftotext first - Fastest and most reliable for plain text

    run_shell command="pdftotext -layout /path/to/file.pdf -"
    
  2. If pdftotext unavailable, check for Python libraries

    run_shell command="python3 -c \"import fitz; print('PyMuPDF available')\""
    
  3. For table/structured data, use pdfplumber

    run_shell command="python3 -c \"import pdfplumber; ...\""
    
  4. Verify extraction succeeded - Check output contains readable text, not binary data

Installation Notes

If tools are not available:

# Ubuntu/Debian
apt-get install poppler-utils  # pdftotext, pdfinfo

# Install Python libraries
pip install pymupdf pdfplumber

Anti-Patterns to Avoid

  • Do NOT rely solely on read_file with filetype='pdf' for critical text extraction
  • Do NOT assume PDF text is in any particular order - verify extracted content
  • Do NOT use image-based extraction unless the PDF is scanned (use OCR instead)

Example: Complete Extraction Pattern

# Step 1: Try pdftotext
RESULT=$(pdftotext -layout /path/to/form.pdf - 2>/dev/null)

# Step 2: Verify we got text, not error
if [ -z "$RESULT" ]; then
    # Fallback to Python
    RESULT=$(python3 -c "import fitz; doc=fitz.open('/path/to/form.pdf'); print(''.join(p.get_text() for p in doc))")
fi

# Step 3: Use extracted text
echo "$RESULT"

04 同仓库更多 Skills

HKUDS
HKUDS/AgentPilot

agentpilot

Set up AgentPilot and use its single entry to prepare missing inputs and run configured RSI experiments. Use when the user requests AgentPilot installation or work with AgentPilot.

★ 8.0k
HKUDS
HKUDS/AgentPilot

remember

Review auto-memory entries and propose promotions to OPENSPACE.md, OPENSPACE.local.md, or retained auto-memory topics; detect duplicates, stale facts, conflicts, and ambiguous entries before making changes.

★ 8.0k
HKUDS
HKUDS/AgentPilot

skill-discovery

Search reusable OpenSpace skills locally, and browse cloud packages step by step when a cloud skill may help. Reusing proven skills saves tokens, improves reliability, and extends your capabilities beyond built-in tools.

Skill与智能体开发
★ 8.0k
HKUDS
HKUDS/AgentPilot

delegate-task

Delegate tasks to OpenSpace — a full-stack autonomous worker for coding, DevOps, web research, and desktop automation, backed by an extensive MCP tool and skill library. Skills auto-improve through use, reducing token consumption over time. A cloud community lets agents share and collectively evolve reusable skills.

Skill与智能体开发
★ 8.0k
HKUDS
HKUDS/AgentPilot

write-special-chars-ts-check

Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, plus a reliable TypeScript type-checking step that filters output to only the newly-written files to avoid being misled by pre-existing errors.

★ 8.0k
HKUDS
HKUDS/AgentPilot

write-special-chars-readback

Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, with mandatory post-write verification via read_file and cleanup of /tmp helper scripts.

★ 8.0k