Extract cited content from large documents via outline and targeted section reads.
npx skills add https://github.com/wmorein/docling-reading-skill --skill docling-reader
Please help me install this Agent Skill. Skill: docling-reader Source: https://github.com/wmorein/docling-reading-skill/tree/main Command: npx skills add https://github.com/wmorein/docling-reading-skill --skill docling-reader
Large documents are expensive to scan in full; this skill enables intelligent partial reading so you load only the sections you need, saving time and reducing context load.
Command it to read the Introduction of a 120k-character document and return a concise summary with citations.
High-intent search queries and answers about installing and using this skill.
To extract specific sections from a large document, use an outline-first reading strategy to fetch only targeted content. This approach loads necessary sections from plain text or markdown, saving time and reducing context load.
A docling JSON sidecar provides structural metadata for document reading. It enables targeted section and line-based retrieval, allowing you to perform outline-first analysis and extract precise content with citations without scanning the full document.
Yes, partial reading strategies support plain text and markdown files. The skill applies a size and structure-based reading strategy to these formats to deliver precise results with citations to source sections.
To summarize a 120k-character document without full scanning, command a targeted section read. The skill analyzes the outline and retrieves only the requested sections, returning a concise summary with citations.
The best way to retrieve cited content from large documents is using targeted section and line-based extraction. This method provides contextual extraction with citations to source sections, ensuring precise answers without full text loads.
Use partial reading for large documents when full scanning is too expensive. It enables intelligent outline-first analysis and targeted retrieval, loading only necessary sections to reduce context load and save time.
takusaotomeConvert PDFs, office documents, and images into Markdown and structured formats.
vincentchanAnalyze content from URLs and compile analyses into a swipe file.
orbrunoExtract structure-aware chunks with page and heading metadata for RAG workflows.
aiyinluyaSearch academic papers across ArXiv, Semantic Scholar, PubMed, and Wikipedia.
