Convert Korean HWP-family documents to Markdown for downstream processing.
npx skills add https://github.com/gracefullight/krds --skill oma-hwp-gracefullight
Please help me install this Agent Skill. Skill: oma-hwp Source: https://github.com/gracefullight/krds/tree/main/.agents/skills/oma-hwp Command: npx skills add https://github.com/gracefullight/krds --skill oma-hwp-gracefullight
It solves the problem of turning Korean HWP-family documents into AI- and RAG-ready content by converting HWP/HWPX/HWPML into clean Markdown while preserving headings, tables, lists, images, footnotes, and hyperlinks.
.hwp, .hwpx, and .hwpml into readable Markdown with extracted document structure for downstream analysis.Convert an attached Korean HWP file to Markdown and place the output next to the source file by telling the assistant: "Convert input_path to Markdown output_path using oma-hwp."
High-intent search queries and answers about installing and using this skill.
You convert HWP files to Markdown for RAG ingestion by using a conversion tool that extracts text, headings, tables, and lists from Korean documents while stripping invalid Hancom Private Use Area characters for clean downstream processing.
Yes, you can extract tables from HWP documents into Markdown. The conversion process flattens merged-cell tables into GFM-friendly formats, ensuring structured table data remains usable for downstream reading and LLM workflows.
The best way to parse HWPX files for LLM workflows is using a dedicated Korean document parser that converts HWP-family formats into clean Markdown, preserving structural elements like footnotes and hyperlinks for AI-ready content.
Yes, you need Bun and Bunx installed locally along with the kordoc library to convert HWPML files to text, as the runtime validates file extensions and executes the extraction process.
Broken characters in HWP to Markdown conversion often occur due to Hancom Private Use Area glyphs that render incorrectly without original fonts, which is why post-processing strips these invalid characters automatically.
ddarkrConvert HWP, HWPX, and HWPML files into Markdown or JSON.
patrykkopycinskiSet up a RAG pipeline with Elasticsearch for vector retrieval.
badlogicAutomate Chrome browser control and content extraction via the Chrome DevTools Protocol.
first-flukeConvert Korean HWP, HWPX, and HWPML files into structured Markdown.