---
name: pdf-extraction-and-summary
description: "Extracts facts, tables, and decisions from a long PDF into a structured summary where every point carries its page reference, so claims can be checked fast. Use when a long PDF needs its key facts, tables, decisions, and deadlines pulled into a structured summary with page references. Do not use for translating a document or editing the original file. Related topics: productivity, research."
---

# PDF Extraction and Summary

Extracts facts, tables, and decisions from a long PDF into a structured summary where every point carries its page reference, so claims can be checked fast.

## When to use

Use this skill when a long PDF needs its key facts, tables, decisions, and deadlines pulled into a structured summary with page references. Do not use it for translating a document or editing the original file.

## Workflow

1. Ask the user for each input listed below in a single message, skipping any they have already given. Where options are listed, offer them, but accept other answers too.
2. Read the full text first and write a short overview giving the document type, author, date, and main conclusion.
3. Answer each stated question directly, placing a page reference after every claim.
4. Extract the facts, figures with units, and the tables that matter, noting unreadable cells instead of guessing.
5. Collect decisions, obligations, deadlines, and open issues in their own section, with owners where named.
6. Flag contradictions and unanswered questions, then list any sections left uncovered.

## Inputs

- **Document text** (longer free text)
- **Document type** (choose one: Contract, Research paper, Financial report, Policy, Manual, Other)
- **What you need from it** (longer free text)
- **Summary length** (choose one: One page, Two pages, Detailed)
- **Extract tables** (choose one: Yes, No)
- **Reader** (free text, e.g. "Who will use the summary")

## Output

A summary at the chosen length with an overview, answers with page references, key figures, tables, a decisions and deadlines section, and a list of gaps.

## Prompt template

Fill the inputs into this template and follow it as the full instruction.

### PDF Extraction and Summary

Pull what matters out of this document and keep the page references.

- Document text: {{Document text}}
- Document type: {{Document type}}
- What you need from it: {{What you need from it}}
- Summary length: {{Summary length}}
- Extract tables: {{Extract tables}}
- Reader: {{Reader}}

Read the whole text before writing anything. Start with a brief overview of what the document is, who wrote it, its date, and its main conclusion. Then answer each stated question directly, with the page number in brackets after every claim, e.g. (p. 14). List the key facts and figures with units, and reproduce the tables that matter as clean tables with the page they came from, noting any row that was unreadable rather than guessing. Collect decisions, obligations, deadlines, and open issues in a separate section with their owners if named. Where the document contradicts itself or leaves a question unanswered, say so plainly.

Keep to the chosen length. Quote the original only when exact wording matters, and mark quotations. Finish with a short list of the sections not covered and why.

## Pitfalls

- Citing the page of the summary heading instead of the page the fact appears on.
- Guessing a number from a garbled table cell, which turns an extraction error into a false fact.
- Smoothing over a contradiction between sections, hiding the very issue the reader needs to know about.