skills.rest
ever-just
ever-justCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

notebook

Build and query document knowledge bases with indexed metadata and citations.

3|1|Updated Apr 8, 2026
Search Tags:#note-taking#research#knowledge base#citations#web scraping#pdf extraction#puppeteer
One-click install
npx skills add https://github.com/ever-just/agentskills --skill notebook-ever-just
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: notebook
Source: https://github.com/ever-just/agentskills/tree/main/skills/white-paper-writing/blastum-notebook
Command: npx skills add https://github.com/ever-just/agentskills --skill notebook-ever-just

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pdfplumber, html2text, requests, puppeteer, puppeteer-extra, puppeteer-extra-plugin-stealth, @ghostery/adblocker-puppeteer, robots-parser, and includes scripts (resource) and references (resource) components.

What problem does it solve? Research scattered across web pages, PDFs, and notes gets lost between sessions, forcing agents to re-fetch sources and re-derive facts every time. This Skill acquires, indexes, and summarizes document collections into persistent topic notebooks so answers come with citations from saved sources. ## Core Features & Use Cases - Source Acquisition: Fetches web pages via headless Chrome (puppeteer with stealth and robots.txt compliance) or Python fallback, downloads and extracts PDFs with page markers, and delegates batches of 3+ sources to parallel subagents. - Indexed Knowledge Base: Writes per-source metadata files, document summaries, and a topic index so queries search small metadata first and only escalate to raw documents when needed. - Persistent Memory: Appends durable facts, decisions, and preferences to notes files, supports global notebooks with checkout/check-in across workspaces, and exports snapshots for NotebookLM import. - Use Case: Researching TIPS ladder retirement strategies — the agent collects 15 sources into a tips-ladders notebook, extracts key facts with citations, and answers follow-up questions weeks later without re-fetching anything. ## Quick Start Ask the agent to research a topic and start a notebook, for example: "Research TIPS ladder strategies and save the sources in a notebook."

Frequently Asked Questions about notebook

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

FAQPage Schema
How do I build a searchable document knowledge base with an AI agent?▼

Create a topic notebook with documents, metadata, and notes directories, then acquire sources via web fetch or PDF download. Each source gets a metadata file and summary, and a topic index is rebuilt so queries search metadata first before reading raw documents.

How to fetch web pages that block bots or require CAPTCHA?▼

Use the puppeteer fetch script with the --interactive flag, which opens a visible browser and waits up to five minutes for manual CAPTCHA solving. Alternatively, launch Chrome with remote debugging and capture the live tab after scrolling content yourself.

Does the notebook skill work with scanned PDFs?▼

Standard extraction via pdfplumber fails on scanned or image-only PDFs. When extracted text averages under 100 characters per page, the workflow falls back to a vision-ocr skill and uses the OCR output as the document content.

Can I share notebooks across multiple projects or workspaces?▼

Yes, global notebooks live under ~/.cursor/notebooks and can be checked out into any workspace, synced back, or checked in using the notebook-library.sh script. Checkout state is tracked so one notebook can exist in multiple workspaces.

What dependencies does the notebook skill require?▼

Python scripts need pdfplumber, html2text, and requests from requirements.txt. Web acquisition needs Node 18+ with puppeteer, puppeteer-extra-plugin-stealth, and an adblocker installed once in the vendor/puppeteer directory.

Related Skills

Explore More Education & Research Skills→
SahilFruitwalaSahilFruitwala

topic-finder

Fetches and dedupes niche headline candidates from RSS feeds and web search for ranking.

Community
Intermediate
jacketlong23jacketlong23

notion-research-documentation

Synthesize information from multiple Notion pages into structured documents with citations.

Community
Advanced
hmzainjamilhmzainjamil
1

PDF OCR Extraction

Extract text from scanned PDFs and image-based documents using OCR.

Community
Intermediate
shalevaminshalevamin

rag-architect

Designs and implements RAG systems with chunking, embedding generation, vector stores, hybrid search, reranking, and evaluation.

Community
Advanced