{
  "$schema": "https://top-agent-skills.com/schemas/catalog.v1.json",
  "type": "primary-category",
  "slug": "document-processing",
  "title": "Document Processing",
  "shortTitle": "Documents",
  "description": "Create, parse, and edit DOCX, PDF, XLSX and PPTX files. OCR, format conversion, deterministic scripts.",
  "canonical": "https://top-agent-skills.com/top-skills-for-document-processing",
  "alternates": {
    "html": "https://top-agent-skills.com/top-skills-for-document-processing",
    "markdown": "https://top-agent-skills.com/top-skills-for-document-processing.md",
    "json": "https://top-agent-skills.com/top-skills-for-document-processing.json"
  },
  "rubricVersion": "1.0",
  "methodology": "https://top-agent-skills.com/about/methodology",
  "contentLastUpdated": "2026-08-31",
  "count": 16,
  "topPick": {
    "slug": "docling",
    "name": "docling",
    "score": 66,
    "tagline": "One skill covering every document format the category exists for — including scanned PDFs and RAG chunking — from the maintainers of the parser itself.",
    "url": "https://top-agent-skills.com/skill/docling"
  },
  "ranking": [
    {
      "rank": 1,
      "slug": "docling",
      "name": "docling",
      "score": 66,
      "tagline": "One skill covering every document format the category exists for — including scanned PDFs and RAG chunking — from the maintainers of the parser itself.",
      "bestFor": "Reading or converting documents the agent cannot open directly, at 65K stars and MIT: scanned PDFs, Office formats, tables, and chunking a corpus for retrieval.",
      "notIdealFor": "Producing polished Office files. Docling is a reader and converter — for generating a styled DOCX, XLSX or PPTX, the Anthropic document skills are the better fit.",
      "provenance": "verified-org",
      "publisher": "Docling project (LF AI & Data)",
      "license": "MIT",
      "compatibility": [
        "claude-code",
        "cursor",
        "codex"
      ],
      "install": "pip install docling && uvx library-skills --claude",
      "url": "https://top-agent-skills.com/skill/docling",
      "json": "https://top-agent-skills.com/skill/docling.json"
    },
    {
      "rank": 2,
      "slug": "azure-document-translation",
      "name": "azure-ai-translation-document-py",
      "score": 59,
      "tagline": "Batch document translation that preserves layout — the part that breaks when you translate a PDF by extracting text and re-flowing it.",
      "bestFor": "Translating a corpus of Word, PDF, Excel or PowerPoint files where the formatting has to survive the round trip.",
      "notIdealFor": "A single short document, or anything that cannot leave your infrastructure — it is a cloud batch service with Azure credentials attached.",
      "provenance": "verified-org",
      "publisher": "Microsoft",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "npx skills add microsoft/skills --skill azure-ai-translation-document-py",
      "url": "https://top-agent-skills.com/skill/azure-document-translation",
      "json": "https://top-agent-skills.com/skill/azure-document-translation.json"
    },
    {
      "rank": 3,
      "slug": "azure-content-understanding",
      "name": "azure-ai-contentunderstanding-py",
      "score": 56,
      "tagline": "Multimodal extraction across documents, images, audio and video through one Azure client — Microsoft-authored, with an explicit trigger list.",
      "bestFor": "Teams already on Azure that need structured extraction from mixed media, not just PDFs — the case where a single-format parser leaves half the corpus unread.",
      "notIdealFor": "Local or offline processing. It teaches an agent to write correct Azure SDK code, so it needs Azure credentials and sends content to the service.",
      "provenance": "verified-org",
      "publisher": "Microsoft",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "npx skills add microsoft/skills --skill azure-ai-contentunderstanding-py",
      "url": "https://top-agent-skills.com/skill/azure-content-understanding",
      "json": "https://top-agent-skills.com/skill/azure-content-understanding.json"
    },
    {
      "rank": 4,
      "slug": "liteparse",
      "name": "liteparse",
      "score": 53,
      "tagline": "Local document extraction with no API key — read the file, not the whole file.",
      "bestFor": "Looking up a figure or table inside a large document without burning context on the parts you do not need.",
      "notIdealFor": "Complex layout reconstruction or scanned-image OCR — the hosted llamaparse skill in the same repo is the heavier option for that.",
      "provenance": "verified-org",
      "publisher": "LlamaIndex",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "npx skills add run-llama/llamaparse-agent-skills --skill liteparse",
      "url": "https://top-agent-skills.com/skill/liteparse",
      "json": "https://top-agent-skills.com/skill/liteparse.json"
    },
    {
      "rank": 5,
      "slug": "box-for-ai",
      "name": "box",
      "score": 53,
      "tagline": "Box as an agent surface, plus the contract-review workflows nothing else in the catalog covers.",
      "bestFor": "Teams whose documents already live in Box, especially contract review and diligence over files stored there.",
      "notIdealFor": "Anyone not on Box — every skill routes through Box auth and its MCP or CLI.",
      "provenance": "verified-org",
      "publisher": "Box",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "npx skills add box/box-for-ai --skill box",
      "url": "https://top-agent-skills.com/skill/box-for-ai",
      "json": "https://top-agent-skills.com/skill/box-for-ai.json"
    },
    {
      "rank": 6,
      "slug": "pdf",
      "name": "pdf",
      "score": 52,
      "tagline": "Extract, fill, merge. PDF manipulation with bundled scripts so Claude does not waste tokens generating them.",
      "bestFor": "Bulk PDF text/table extraction, form filling, merging.",
      "notIdealFor": null,
      "provenance": "anthropic",
      "publisher": "Anthropic",
      "license": "MIT",
      "compatibility": [
        "claude-code",
        "claude-ai",
        "claude-api",
        "cursor",
        "codex"
      ],
      "install": "npx skills add github.com/anthropics/skills --skill pdf",
      "url": "https://top-agent-skills.com/skill/pdf",
      "json": "https://top-agent-skills.com/skill/pdf.json"
    },
    {
      "rank": 7,
      "slug": "docx",
      "name": "docx",
      "score": 48,
      "tagline": "Word documents end-to-end. Create, edit, comment, track-change.",
      "bestFor": "Generating branded Word reports, edit-with-track-changes workflows.",
      "notIdealFor": null,
      "provenance": "anthropic",
      "publisher": "Anthropic",
      "license": "MIT",
      "compatibility": [
        "claude-code",
        "claude-ai",
        "claude-api"
      ],
      "install": "npx skills add github.com/anthropics/skills --skill docx",
      "url": "https://top-agent-skills.com/skill/docx",
      "json": "https://top-agent-skills.com/skill/docx.json"
    },
    {
      "rank": 8,
      "slug": "xlsx",
      "name": "xlsx",
      "score": 48,
      "tagline": "Excel spreadsheets with real formulas, charts, and formatting — not flat CSV exports.",
      "bestFor": "Financial models, structured reports, anything that needs formulas.",
      "notIdealFor": null,
      "provenance": "anthropic",
      "publisher": "Anthropic",
      "license": "MIT",
      "compatibility": [
        "claude-code",
        "claude-ai",
        "claude-api"
      ],
      "install": "npx skills add github.com/anthropics/skills --skill xlsx",
      "url": "https://top-agent-skills.com/skill/xlsx",
      "json": "https://top-agent-skills.com/skill/xlsx.json"
    },
    {
      "rank": 9,
      "slug": "pptx",
      "name": "pptx",
      "score": 48,
      "tagline": "PowerPoint decks with layouts, charts, speaker notes — programmatic from prose.",
      "bestFor": "Generating internal decks at scale, parsing existing decks for content.",
      "notIdealFor": null,
      "provenance": "anthropic",
      "publisher": "Anthropic",
      "license": "MIT",
      "compatibility": [
        "claude-code",
        "claude-ai",
        "claude-api"
      ],
      "install": "npx skills add github.com/anthropics/skills --skill pptx",
      "url": "https://top-agent-skills.com/skill/pptx",
      "json": "https://top-agent-skills.com/skill/pptx.json"
    },
    {
      "rank": 10,
      "slug": "doc-coauthoring",
      "name": "doc-coauthoring",
      "score": 48,
      "tagline": "Multi-author edits with suggestion mode and conflict resolution.",
      "bestFor": "Editing teams where the agent is one of multiple voices on a doc.",
      "notIdealFor": null,
      "provenance": "anthropic",
      "publisher": "Anthropic",
      "license": "MIT",
      "compatibility": [
        "claude-code",
        "claude-ai"
      ],
      "install": "npx skills add github.com/anthropics/skills --skill doc-coauthoring",
      "url": "https://top-agent-skills.com/skill/doc-coauthoring",
      "json": "https://top-agent-skills.com/skill/doc-coauthoring.json"
    },
    {
      "rank": 11,
      "slug": "markdown-exporter",
      "name": "markdown-exporter",
      "score": 48,
      "tagline": "The widest format matrix of any conversion skill we catalogue, and it executes the conversion rather than describing it.",
      "bestFor": "Turning agent-authored Markdown into whatever format the recipient actually needs — a deck, a spreadsheet, a PDF — without hand-writing a converter each time.",
      "notIdealFor": "Reliable auto-triggering. Its description is a bare capability list with no trigger phrasing, so it often needs to be invoked by name rather than firing on its own.",
      "provenance": "community",
      "publisher": "bowenliang123",
      "license": "Apache-2.0",
      "compatibility": [
        "claude-code"
      ],
      "install": "/plugin marketplace add bowenliang123/markdown-exporter",
      "url": "https://top-agent-skills.com/skill/markdown-exporter",
      "json": "https://top-agent-skills.com/skill/markdown-exporter.json"
    },
    {
      "rank": 12,
      "slug": "zotero-skills",
      "name": "zotero-skills",
      "score": 48,
      "tagline": "Programmatic access to the reference manager academia actually uses.",
      "bestFor": "Keeping a Zotero library organised and queryable from the agent instead of through the desktop client.",
      "notIdealFor": "Literature synthesis or writing — this is library operations. Pair it with research-hub or research-co-pilot, both of which sit above it in the workflow.",
      "provenance": "community",
      "publisher": "Wenyu Chiou",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "npx skills add WenyuChiou/zotero-skills --skill zotero-skills",
      "url": "https://top-agent-skills.com/skill/zotero-skills",
      "json": "https://top-agent-skills.com/skill/zotero-skills.json"
    },
    {
      "rank": 13,
      "slug": "claude-healthcare",
      "name": "claude-healthcare",
      "score": 47,
      "tagline": "FHIR to prior-auth to ICD-10 coding: the official payer/provider/pharma toolkit.",
      "bestFor": "Health-tech teams working against real FHIR servers, coding workflows, or trial documentation inside Claude.",
      "notIdealFor": "Anyone needing an OSS license — the repo ships under Anthropic terms, not an open-source license — or consumer health questions.",
      "provenance": "anthropic",
      "publisher": "Anthropic",
      "license": "No OSS license (Anthropic terms)",
      "compatibility": [
        "claude-code"
      ],
      "install": "/plugin marketplace add anthropics/healthcare",
      "url": "https://top-agent-skills.com/skill/claude-healthcare",
      "json": "https://top-agent-skills.com/skill/claude-healthcare.json"
    },
    {
      "rank": 14,
      "slug": "research-co-pilot",
      "name": "research-co-pilot",
      "score": 47,
      "tagline": "The academic writing half nothing else covers: manuscript, grant, peer review, reviewer response.",
      "bestFor": "Graduate students, postdocs and faculty who need the writing and submission stages supported, not just the reading stage.",
      "notIdealFor": "Reference-library management — that is what the Zotero skills are for. Single maintainer, last upstream commit June 2026, and one of the fifteen ships broken: peer-review’s frontmatter puts an unquoted colon inside its description, so the YAML fails to parse and installers skip that skill entirely.",
      "provenance": "community",
      "publisher": "Marazii",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "npx skills add Marazii/research-co-pilot --skill literature-review",
      "url": "https://top-agent-skills.com/skill/research-co-pilot",
      "json": "https://top-agent-skills.com/skill/research-co-pilot.json"
    },
    {
      "rank": 15,
      "slug": "carbone-templates",
      "name": "carbone",
      "score": 46,
      "tagline": "Data-driven document generation: one template, a thousand filled documents.",
      "bestFor": "Invoices, contracts, certificates, and reports produced at volume from structured data and a designed template.",
      "notIdealFor": "One-off documents — authoring a single file is what the Anthropic docx and xlsx skills are for.",
      "provenance": "verified-org",
      "publisher": "Carbone.io",
      "license": "Apache-2.0",
      "compatibility": [
        "claude-code"
      ],
      "install": "/plugin marketplace add carboneio/carbone-skill",
      "url": "https://top-agent-skills.com/skill/carbone-templates",
      "json": "https://top-agent-skills.com/skill/carbone-templates.json"
    },
    {
      "rank": 16,
      "slug": "research-hub",
      "name": "research-workspace",
      "score": 44,
      "tagline": "Zotero plus Obsidian plus NotebookLM as one agent-operable research workspace.",
      "bestFor": "Researchers with a reference library and a note vault who want triage, synthesis, and topic selection driven from the agent rather than by hand.",
      "notIdealFor": "Anyone not already using Zotero or Obsidian — the value is in orchestrating those tools, not replacing them. Shares a maintainer with the zotero-skills entry.",
      "provenance": "community",
      "publisher": "Wenyu Chiou",
      "license": "MIT",
      "compatibility": [
        "claude-code"
      ],
      "install": "claude plugin marketplace add WenyuChiou/ai-research-skills",
      "url": "https://top-agent-skills.com/skill/research-hub",
      "json": "https://top-agent-skills.com/skill/research-hub.json"
    }
  ],
  "faqs": [
    {
      "question": "What is a document-processing Skill?",
      "answer": "A document-processing Skill is a SKILL.md-format bundle that teaches an AI agent how to read, edit, or generate a specific file type — DOCX, PDF, XLSX, PPTX, CSV — using deterministic scripts rather than ad-hoc code. It ships as a folder the agent loads on demand."
    },
    {
      "question": "Which agents support document Skills today?",
      "answer": "Anthropic's docx/pdf/pptx/xlsx Skills run natively in Claude Code, Claude.ai (Pro/Max/Team/Enterprise), and the Claude API. Codex CLI, Cursor, Gemini CLI and Antigravity all support the SKILL.md spec, with Antigravity bundling 1,000+ pre-installed."
    },
    {
      "question": "Do document Skills work offline?",
      "answer": "Yes for the file manipulation itself — the bundled scripts run locally inside the agent's sandbox. OCR Skills that call cloud vision APIs (Anthropic's file-search, Google Vision) require network access; pure pdfplumber/python-docx flows do not."
    },
    {
      "question": "How do I pick between docx, pdf, pptx and xlsx Skills?",
      "answer": "Pick by output format, not input. If you need to produce a Word file, use docx. If you need a styled deck, use pptx. xlsx Skills handle formulas and pivot tables that generic CSV writers cannot. PDF skills handle both reading (extraction) and writing (form fill, watermark, encrypt)."
    },
    {
      "question": "Are these Skills better than just asking the model to write a script?",
      "answer": "For one-off tasks, prompting works. For anything reproducible — same input, same output, every time — a Skill is meaningfully more reliable. It bundles tested scripts, handles format edge cases the model often gets wrong (encrypted PDFs, .docx tracked changes, .xlsx merged cells), and avoids prompt-to-prompt drift."
    }
  ],
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "attribution": "Top Agent Skills (https://top-agent-skills.com)"
}