Parse messy files
Find skills for parsing PDFs, extracting tables, running OCR, converting documents, and preparing file content for agent workflows.
Try this task
I need my agent to read PDFs, extract tables, and turn documents into structured data.
Agent should be able to
Resolve
Returns the best skill, alternatives, install handoff, risk summary, and safety gate.
Text plan
Plain text version for Codex, Claude Code, Cursor, and custom agent runtimes.
Browse
Open the filtered registry view for this workflow and compare candidates manually.
Workflow map
Extract tables from PDFs
Convert files to markdown
Run OCR over scans
Normalize document metadata
Best first installs
Convert PDFs, Office documents, and web files into clean markdown for agents.
Install
Human review before install
Risk
Safe to try
Agent fit
Claude Code + OpenAI Agents
Updated
Jun 1, 2026
$ npx skills add microsoft/markitdownCreate original visual art, posters, PNG assets, and PDF documents through a clear design philosophy.
Install
Agent install candidate
Risk
Safe to try
Agent fit
Claude Code + OpenAI Agents
Updated
Jul 17, 2026
$ npx skills add anthropics/skills --skill canvas-designTurn websites into clean markdown or structured data for retrieval and agents.
Install
Human review before install
Risk
Safe to try
Agent fit
Claude Code + OpenAI Agents
Updated
Jun 1, 2026
$ npx skills add mendableai/firecrawlSkill shortlist
Review portfolio-company financial packages, flag variance to plan, and prepare monitored KPI summaries.
Review before production · Require human approval before installing into a real workspace.
Automate a real browser from the terminal for navigation, form filling, snapshots, screenshots, extraction, and UI-flow debugging.
Low metadata risk · Require human approval before installing into a real workspace.
Open-source LLM-friendly web crawler and scraper for agent workflows.
Review before production · Review the audit page, then allow agent install in a sandboxed workflow.
Data framework for building RAG and knowledge workflows around agent tasks.
Review before production · Review the audit page, then allow agent install in a sandboxed workflow.
Guidance for distinctive, intentional UI design, typography, visual direction, and non-template-like product interfaces.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Use Playwright to interact with and test local web applications, capture screenshots, debug UI behavior, and inspect browser logs.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Design and implementation guidance for distinctive landing pages, portfolios, product demos, and purposeful redesigns.
Low metadata risk · Allow agent install in a sandbox or low-risk workspace, then promote after one successful narrow task.
Apply Anthropic official brand colors, typography, and visual standards to appropriate Anthropic-related artifacts.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
React and Next.js performance guidance for writing, reviewing, and refactoring production UI code.
Low metadata risk · Allow agent install in a sandbox or low-risk workspace, then promote after one successful narrow task.
Review UI code for web interface guidelines, UX quality, accessibility, and interaction design best practices.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Deploy applications and websites to Vercel, including preview deployments and shareable deployment links.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Translate Figma designs into production-ready UI code with close visual fidelity and component-level detail.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Workflow automation for connecting agents to repeated operational tasks.
Review before production · Require human approval before installing into a real workspace.
Production-grade engineering skills and quality gates for AI coding agents.
Review before production · Require human approval before installing into a real workspace.
Browser automation layer for agents that need to interact with websites.
Review before production · Require human approval before installing into a real workspace.
FAQ
These answers are written for both human builders and agents consuming the Registry API.
Start by comparing MarkItDown, Canvas Design, Firecrawl. OpenAgentSkill ranks them by workflow fit, GitHub adoption, trust score, safety gate, and install readiness.
Yes. Use the linked Registry API prompt to query /api/skills/search with the task: "I need my agent to read PDFs, extract tables, and turn documents into structured data." and retrieve install handoff links for the top results.
No. Start with the highest-fit skill, test it in a sandbox workflow, and add companion skills only when the task needs extra coverage.
Parse messy files
Find skills for parsing PDFs, extracting tables, running OCR, converting documents, and preparing file content for agent workflows.
Try this task
I need my agent to read PDFs, extract tables, and turn documents into structured data.
Agent should be able to
Resolve
Returns the best skill, alternatives, install handoff, risk summary, and safety gate.
Text plan
Plain text version for Codex, Claude Code, Cursor, and custom agent runtimes.
Browse
Open the filtered registry view for this workflow and compare candidates manually.
Workflow map
Extract tables from PDFs
Convert files to markdown
Run OCR over scans
Normalize document metadata
Best first installs
Convert PDFs, Office documents, and web files into clean markdown for agents.
Install
Human review before install
Risk
Safe to try
Agent fit
Claude Code + OpenAI Agents
Updated
Jun 1, 2026
$ npx skills add microsoft/markitdownCreate original visual art, posters, PNG assets, and PDF documents through a clear design philosophy.
Install
Agent install candidate
Risk
Safe to try
Agent fit
Claude Code + OpenAI Agents
Updated
Jul 17, 2026
$ npx skills add anthropics/skills --skill canvas-designTurn websites into clean markdown or structured data for retrieval and agents.
Install
Human review before install
Risk
Safe to try
Agent fit
Claude Code + OpenAI Agents
Updated
Jun 1, 2026
$ npx skills add mendableai/firecrawlSkill shortlist
Review portfolio-company financial packages, flag variance to plan, and prepare monitored KPI summaries.
Review before production · Require human approval before installing into a real workspace.
Automate a real browser from the terminal for navigation, form filling, snapshots, screenshots, extraction, and UI-flow debugging.
Low metadata risk · Require human approval before installing into a real workspace.
Open-source LLM-friendly web crawler and scraper for agent workflows.
Review before production · Review the audit page, then allow agent install in a sandboxed workflow.
Data framework for building RAG and knowledge workflows around agent tasks.
Review before production · Review the audit page, then allow agent install in a sandboxed workflow.
Guidance for distinctive, intentional UI design, typography, visual direction, and non-template-like product interfaces.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Use Playwright to interact with and test local web applications, capture screenshots, debug UI behavior, and inspect browser logs.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Design and implementation guidance for distinctive landing pages, portfolios, product demos, and purposeful redesigns.
Low metadata risk · Allow agent install in a sandbox or low-risk workspace, then promote after one successful narrow task.
Apply Anthropic official brand colors, typography, and visual standards to appropriate Anthropic-related artifacts.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
React and Next.js performance guidance for writing, reviewing, and refactoring production UI code.
Low metadata risk · Allow agent install in a sandbox or low-risk workspace, then promote after one successful narrow task.
Review UI code for web interface guidelines, UX quality, accessibility, and interaction design best practices.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Deploy applications and websites to Vercel, including preview deployments and shareable deployment links.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Translate Figma designs into production-ready UI code with close visual fidelity and component-level detail.
Low metadata risk · Review the audit page, then allow agent install in a sandboxed workflow.
Workflow automation for connecting agents to repeated operational tasks.
Review before production · Require human approval before installing into a real workspace.
Production-grade engineering skills and quality gates for AI coding agents.
Review before production · Require human approval before installing into a real workspace.
Browser automation layer for agents that need to interact with websites.
Review before production · Require human approval before installing into a real workspace.
FAQ
These answers are written for both human builders and agents consuming the Registry API.
Start by comparing MarkItDown, Canvas Design, Firecrawl. OpenAgentSkill ranks them by workflow fit, GitHub adoption, trust score, safety gate, and install readiness.
Yes. Use the linked Registry API prompt to query /api/skills/search with the task: "I need my agent to read PDFs, extract tables, and turn documents into structured data." and retrieve install handoff links for the top results.
No. Start with the highest-fit skill, test it in a sandbox workflow, and add companion skills only when the task needs extra coverage.