# PDF

> Read, extract, summarize, and analyze PDF attachments saved by Kivio Chat.

- Skill: `zmgid/pdf` (Agent Skill)
- Install (CLI): `npx skillmds@latest add zmgid/pdf`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zmgid/pdf/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: zmgid (https://skillmd.com/u/zmgid)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/zmgid/pdf

---


# PDF Skill

Use this skill when the user attaches or references a PDF and asks to read, summarize, extract, compare, translate, inspect, or answer questions about it.

## Inputs

Kivio stores each uploaded file as a safe local copy and includes its absolute path in the user message under `Kivio 安全副本路径`. Use that host path with `bash`. You can also process any local PDF discovered via `glob` / `read` (directory listing).

`read` does not parse binary PDFs.

## Workflow

1. Identify the PDF path from the attachment note or a local search.
2. If the host has a PDF tool (`pdftotext`, `python` with pypdf, OfficeCLI, etc.), extract text with `bash`. For a multi-line script, `write` it first, then run it — do not cram quoted code into `python -c`.
3. If no PDF tool is installed, say so and stop guessing. Do not invent document contents from the filename.
4. If extraction returns little text, the PDF may be scanned/image-only — ask for OCR or use Lens.
5. For long PDFs, extract page-level text first, then summarize by section/page before answering.

## Output

- For summaries: include the main points and mention any pages/sections you used when available.
- For extraction: preserve original order and tables/lists as much as practical.
- For analysis: quote only short snippets and explain conclusions separately.