AgentSkillsCN

pdf-summary

通过 Python（pymupdf）提取文本，生成结构化摘要，从而对 PDF 文档进行精简概括。当用户希望从 PDF 文件中获取摘要、概览或关键要点时，这套工具将为您省去大量阅读与整理的时间。

SKILL.md
--- frontmatter
name: pdf-summary
description: Summarize PDF documents by extracting text with Python (pymupdf) and generating structured summaries. Use when the user wants a summary, overview, or key points from a PDF file.
compatibility: Requires Python 3 and pymupdf (pip install pymupdf)

PDF Summary

Extract text from PDF files and produce structured summaries.

When to use this skill

  • •User asks to summarize a PDF document
  • •User wants key points or an executive overview of a PDF
  • •User needs a chapter-by-chapter breakdown of a long PDF

Steps

  1. •

    Ensure pymupdf is installed. If not:

    bash
    pip install pymupdf
  2. •

    Extract text from the PDF using the bundled script.

    bash
    python scripts/extract_text.py "INPUT_FILE_PATH"

    The script prints extracted text directly to stdout (page-separated). You do NOT need to read a separate file — just use the shell output. Options: --quiet (no page markers), --save (also write a .txt file), --password <pwd> (encrypted PDF).

  3. •

    Produce a summary based on the extracted text in this structure:

    • •Document: title or filename
    • •Pages: total count
    • •Executive Summary: 2–3 sentence overview
    • •Key Points: 5–10 bullet items
    • •Detailed Sections: section-by-section breakdown (if requested)

Edge cases

  • •Scanned/image-only PDFs: If extracted text is empty or garbled, inform the user that OCR is required and suggest pip install pytesseract.
  • •Very large PDFs (100+ pages): Extract text first, then summarize in chunks rather than loading the entire text at once.
  • •Password-protected PDFs: pymupdf supports passwords — pass it via fitz.open(path, password="xxx"). Ask the user for the password if needed.

Scripts