skills.rest
petobens
petobensCOMMUNITY
Explore all skills by this author →
GitHub

read-pdf

Reads local and Google Drive PDFs and answers questions about their contents.

Updated Mar 29, 2024
Search Tags:#pdf#read#document#extract#summarize#google drive#pdftotext
One-click install
npx skills add https://github.com/petobens/ai-harness --skill read-pdf-petobens
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: read-pdf
Source: https://github.com/petobens/ai-harness/tree/main/skills/read-pdf
Command: npx skills add https://github.com/petobens/ai-harness --skill read-pdf-petobens

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reading a PDF inside an AI agent is inconsistent: some agents parse PDFs natively while others only see raw bytes, and PDFs stored in Google Drive must be downloaded first. This Skill provides a single workflow that detects the running agent, picks the right reading path, and handles both local files and Drive-hosted documents. ## Core Features & Use Cases - Agent-aware reading: Detects whether Claude Code or Codex is running and uses the native PDF reader or CLI tools (pdfinfo, pdftotext, pdftoppm) accordingly. - Google Drive support: Downloads Drive PDFs by file ID using gws, delegating file discovery to the google-drive skill, then reads the local copy. - Scanned PDF fallback: Renders image-only pages to PNGs at 150 DPI for visual inspection when no text layer exists. - Use Case: A user shares a Google Drive link to a 60-page contract and asks for the termination clause. The Skill downloads the PDF, reads it in page windows, and answers with the relevant pages cited. ## Quick Start Read the PDF at ./reports/q3.pdf and tell me how many pages it has and what it covers.

Frequently Asked Questions about read-pdf

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I read a PDF file with an AI agent?▼

Detect which agent is running first: Claude Code reads PDFs natively with its built-in Read tool, while Codex and other agents use CLI tools. Run pdfinfo for page count, then pdftotext -layout to extract the text layer to stdout.

How to read a PDF stored in Google Drive?▼

Download the Drive PDF first using gws drive files get with the bare file ID and alt=media, saving to a relative path in the current directory. Then read the downloaded local copy with the agent-appropriate method.

Can I extract text from a scanned or image-only PDF?▼

Scanned PDFs have no text layer, so pdftotext returns little or nothing. Render pages to PNGs at 150 DPI with pdftoppm and view them if the agent accepts image input; otherwise no OCR is available and the limitation is reported.

Why does reading a PDF return raw bytes instead of text?▼

This happens when an agent without a native PDF parser, such as Codex, points its file read at a PDF. Use the CLI path instead: pdftotext -layout extracts the text layer, with -f and -l flags for page windows.

How do I read a long PDF without overloading context?▼

Read long PDFs in page windows rather than all at once. Claude Code accepts a pages range like 1-5 with a maximum of about 20 pages per read, and pdftotext supports -f and -l flags to bound the extraction.

Related Skills

Explore More Personal & Entrepreneur Skills→
takimoto-sketchtakimoto-sketch

gdrive

Search, read, and download Google Drive files with folder scoping.

Community
Advanced
ddarkrddarkr
5

oma-pdf

Convert PDF files into structured Markdown with text, tables, headings, and images.

Community
Intermediate
jinming99jinming99
1

skill-1c-concern-extraction-qc

Cross-verify concern sheet YAML against review and paper PDFs.

Community
Intermediate
OpenSenseNovaOpenSenseNova
5.3k

pdf-analysis

Extract text, tables, and embedded images from text-based and scanned PDF documents.

Official
Intermediate