---
name: pdf-summarizer
description: Summarizes PDF documents and extracts key points as a structured list.
  Use when asked to "summarize pdf", "extract key points", "read this document",
  "what does this pdf say", or "summarize this file". Accepts a file path or URL
  to a PDF. Returns a structured summary with title, key points, and word count.
---

This skill reads a PDF from a local path or URL, extracts the full text content,
and produces a structured summary with the document title, a concise overview,
and a list of the most important points. It handles multi-page documents, skips
headers and footers, and falls back gracefully if the PDF is image-based.

## When to use this skill
- User uploads or references a PDF file and asks for a summary
- User asks "what are the key points in this document"
- User wants a quick overview before reading a long report

## Input requirements
- file_path: string (required) — local path or HTTPS URL to PDF file
- max_points: number (optional) — maximum key points to extract, default 5
- language: string (optional) — output language, default "en"

## Output format
Returns JSON object:
{
  "title": string,
  "page_count": number,
  "word_count": number,
  "summary": string,
  "key_points": ["point 1", "point 2", ...],
  "error": string | null
}

## Error handling
- File not found → return error with path validation message
- Image-based PDF (no text) → return error with OCR suggestion
- Corrupt file → return error with file validation message
- Network timeout (URL) → retry once, then return error

## Permission scope
network_access: true   # needed for URL-based PDFs
file_access: true      # needed for local file paths
system_commands: false

## Dependencies
pip install pdfplumber requests

## Examples
User: "Summarize this PDF: /documents/report.pdf"
User: "What are the key points in https://example.com/whitepaper.pdf"