#01
MarkItDown
Convert PDFs, Office documents, and web files into clean markdown for agents.
100
Quality
79
Trust
—
Proven
Document ProcessingJun 1, 2026 pushMITNeeds first agent run
$ npx skills add microsoft/markitdownDocument skills
PDF extraction skills for AI agents.
Compare skills for PDF parsing, OCR, table extraction, markdown conversion, document metadata, and agent-ready file processing.
Built for users searching for AI agent skills that can parse PDFs, extract tables, and convert documents into usable context.
Matched
5
Stars
385K
Input
Output
Markdown
Agent jobs
Start from a real workflow, not a keyword.
These pages are built for high-intent search and for agents that need a structured shortlist with install commands, trust signals, audit links, and real outcome evidence before installing third-party code.
01
Convert PDFs into clean markdown for agents
02
Extract tables and metadata from reports
03
Prepare legal, finance, and research documents for review
04
Use OCR fallback when scanned pages need text extraction
Task routes
Task pages agents can call.
Ranked shortlist
High-signal skills to inspect first.
80K stars
163K stars
Create original visual art, posters, PNG assets, and PDF documents through a clear design philosophy.
100
Quality
97
Trust
77
Proven
design-creativeJul 17, 2026 pushSource terms (see LICENSE.txt)Promising agent evidence
$ npx skills add anthropics/skills --skill canvas-design#03
Firecrawl
Turn websites into clean markdown or structured data for retrieval and agents.
100
Quality
80
Trust
—
Proven
Web ScrapingJun 1, 2026 pushAGPL-3.0Needs first agent run
$ npx skills add mendableai/firecrawl#04
Crawl4AI
Open-source LLM-friendly web crawler and scraper for agent workflows.
100
Quality
79
Trust
39
Proven
Web ScrapingJun 1, 2026 pushApache-2.0Early agent signal
$ npx skills add unclecode/crawl4ai#05
LlamaIndex
Data framework for building RAG and knowledge workflows around agent tasks.
100
Quality
79
Trust
—
Proven
RAGJun 1, 2026 pushMITNeeds first agent run
$ npx skills add run-llama/llama_indexEvaluation
How to choose the right skill.
Handles layout, headings, and tables without destroying context
Reports extraction limits and OCR uncertainty
Supports batch or repeatable processing
Documents privacy and local processing assumptions
Questions
Which PDF skill should I choose first?
Choose a skill that supports your document type, preserves tables or headings, and makes extraction failures visible instead of silently guessing.
Can these skills handle scanned PDFs?
Some can, but scanned PDFs usually need OCR and human review for high-stakes documents.