Agent skill · software engineering · aradotso
kordoc-korean-document-parser
Parse HWP, HWPX, and PDF Korean documents to Markdown using kordoc — supports CLI, programmatic API, and MCP server integration.
Why this skill is useful
Adds executable scripts for parsing proprietary Korean document formats into Markdown and structured data, which the AI cannot generate on its own.
What it needs
Requires kordoc, pdfjs-dist installed locally. About 5k tokens when loaded. Last updated 2026-07-09. 70 stars on the source repository.
What this skill does
kordoc Korean Document Parser Skill by ara.so — Daily 2026 Skills collection. kordoc is a TypeScript library and CLI for parsing Korean government documents (HWP 5.x, HWPX, PDF) into Markdown and structured IRBlock[] data. It handles proprietary HWP binary formats, table extraction, form field recognition, document diffing, and reverse Markdown→HWPX generation. --- Installation --- Core API Auto-detect and Parse Any Document Format-Specific Parsers Parse Options --- Working with IRBlocks Convert Blocks Back to Markdown --- Document Comparison --- Form Field Extraction --- Markdown → HWPX Generation --- CLI Usage --- MCP Server Setup Add to your MCP config (Claude Desktop, Cursor, Windsurf): Available MCP Tools Tool Description ------ ------------- parsedocument Parse HWP/HWPX/PDF → Markdown + metadata + outline + warnings detectformat Detect file format via magic bytes parsemetadata Extract only metadata (fast, no full parse) parsepages Parse a specific page range parsetable Extract the Nth table from a document comparedocuments Diff two documents (cross-format supported) parseform Extract form fields as structured JSON --- TypeScript Types Reference --- Common Patterns Batch Process Files with Error Handling Extract All Tables from a Document OCR with Tesseract.js Watch Mode Programmatic API --- Troubleshooting buffer.buffer vs Buffer — kordoc requires ArrayBuffer, not Node.js Buffer. Always pass readFileSync("file").buffer or use .buffer on a Uint8Array. PDF tables not detected — Line-based detection requires pdfjs-dist installed. Install it: npm install pdfjs-dist. For borderless tables, kordoc uses cluster-based heuristics automatically. "IMAGEBASEDPDF" error — The PDF contains scanned images with no text layer. Provide an ocr function in parse options. "ENCRYPTED" error — HWP DRM/password-protected files cannot be parsed without the decryption key. No workaround. Korean characters garbled in output — Ensure your terminal/file uses UTF-8 encoding. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aradotso/kordoc-korean-document-parser