Extracts text from Word documents into Markdown clippings for an Obsidian wiki.
npx skills add https://github.com/ehartye/wiki-master --skill clip-docx-ehartye
Please help me install this Agent Skill. Skill: clip-docx Source: https://github.com/ehartye/wiki-master/tree/main/skills/clip-docx Command: npx skills add https://github.com/ehartye/wiki-master --skill clip-docx-ehartye
--wrap=none to preserve verbatim spans across paragraphs. - Duplicate and decline tracking: Skips already-clipped documents, records declines for thin or failed extractions, and supports --allow-short for genuinely brief documents. - Research-run provenance: Records source URL, quality tier, content hash, and an optional --topic so /wiki-triage can group clips from the same discovery run. - Use Case: You download an Academia.edu paper served as a .docx file. Clip it into raw/clippings/ with its DOI as the source, then hand off to /wiki-ingest for summarization and cross-referencing. ## Quick Start Clip the file paper.docx into the wiki with its DOI as the source and the current research topic attached.High-intent search queries and answers about installing and using this skill.
Run the clip-docx script with the document path, an optional source URL, quality tier, and topic. It uses pandoc to extract text and writes a Markdown clipping with standard frontmatter into raw/clippings/, ready for ingestion.
Pandoc extracts text from .docx files by reading the document XML directly, producing clean Unicode prose. This Skill invokes pandoc with --wrap=none and -t plain to keep paragraphs intact for quoting.
No. Corrupt or password-protected documents fail fast and are reported for manual handling rather than fabricated. Unlike PDF clipping, there is no OCR path because pandoc reads the document XML directly.
Documents under 100 words are flagged as thin to catch failed extractions. If the document is genuinely brief, such as a one-page handout, verify the extraction is complete and re-clip with the --allow-short flag.
No. Clippings under raw/ are immutable source-of-truth; only the clipping script writes there. Corrections belong in downstream wiki pages, and quotes should be verified against the original document.
giraphantDownload academic books via MD5 and papers via DOI or URL.
Strategic-AutomationIngest sources into an interlinked Markdown wiki with linting and queries.
jonnytacos8Summarize key figures and trends from SharePoint Excel tables using rag_search.
thimsluggaGenerate importable Obsidian Web Clipper JSON templates from a target page and Base definition.