القائمة
Convert DOCX files into markdown while preserving headings, lists, tables, metadata, and extracted images.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Interview-based vault context capture for cmds-vault. Personalizes BRAIN.md, fills 100 Themes stubs, and optionally fills the CMDS.md Vault Operator profile so AI assistants give context-aware answers from day one. Activate when the user says "온보딩해줘", "처음 시작할게", "cmds onboarding", "온보딩", "context onboarding", "내 컨텍스트로 채워줘", "fill my vault", "personalize my brain", "interview me", "어떻게 시작하지", or after a fresh clone before exploring slash commands. Complementary to gobi-onboarding (which focuses on Gobi Space publishing).
Export Obsidian Canvas (.canvas) files to JPG images using Playwright. Use when user asks to export, screenshot, or convert a canvas to an image.
Help organize thoughts through question-centered dialogue. Acts as a questioning thinking partner rather than an answer-providing assistant. Use when user needs to clarify problems, explore decisions, brainstorm strategy, or reflect on their own thinking.
Generate or modify a Gobi Brain homepage (home.html) using one of 4 style templates (neon-terminal, minimal-editorial, magazine, brutalist) with an interview that customizes hero links, gobi.* features to include, accent color, and fonts. Use when asked to create a Brain homepage, redesign home.html, build a Gobi vault landing page, or apply a different style to an existing Brain page.
Convert Obsidian markdown documents (with Mermaid diagrams, tables, wiki links) to professional A4 PDF using Playwright. Use when user requests PDF export of markdown files.
Create and manage Obsidian Canvas files with automatic layout generation. Use when creating visual knowledge maps, weekly reading summaries, or project timelines.
| name | docx-to-markdown |
| description | Convert DOCX files into markdown while preserving headings, lists, tables, metadata, and extracted images. |
Convert DOCX files to well-formatted markdown files with images extracted.
python3 docx_to_markdown.py "<docx_path>" -o "<output_path>"
_files_/ folder with document prefix20. Literature Notes/Books/ # CMDS literature folder (typical for books)
├── document.docx # Original file
├── document.md # Extracted markdown
└── _files_/ # Images folder (sibling to .md)
├── DocTitle_image1.png
├── DocTitle_figure2.jpg
└── ...
Pick the destination based on source type:
| Source type | Recommended -o path |
|---|---|
| Book / long-form | 20. Literature Notes/Books/<title>.md |
| Article / paper | 20. Literature Notes/Articles/<title>.md |
| Apple Notes export | 20. Literature Notes/AppleNotes/<title>.md |
| Triage (unsure) | 00. Inbox/02. Clippings/<title>.md, then promote later |
Images extract to a _files_/ sibling folder relative to the output .md. After conversion, link images via the relative  markdown the script emits — Obsidian resolves them automatically.
Pairs with: /connect (capture as theme stub) or cmds-llm-wiki-ingest (compile into LLM Wiki) once the .md is in place.
Images are extracted with a document prefix derived from the title:
DieEmptyUnleash (first 3 words, special chars removed)_files_/DieEmptyUnleash_image1.pngThis prevents filename collisions when extracting multiple documents to the same folder.
The script emits CMDS-conformant frontmatter (per .claude/rules/frontmatter-standard.md) plus optional provenance fields:
---
type: literature
aliases: []
description: "Markdown extracted from DOCX: original.docx. {subject if present}"
author:
- "[[Me]]"
date created: YYYY-MM-DD
date modified: YYYY-MM-DD
tags:
- literature
- imported
status: unread
source_file: "original.docx"
source_type: docx
extracted: YYYY-MM-DD HH:MM:SS
---
# {Document Title}
**Source author**: {DOCX core_properties author, if present}
{content with  links}
Notes:
author: [[Me]] is the CMDS placeholder for user-created notes (resolved by the WELCOME ritual). The original DOCX author is surfaced as **Source author**: in the body so it stays visible without producing a YAML wikilink orphan.status: unread follows the CMDS 5-value enum (unread/reading/inProgress/completed/archived).type: (e.g. to note, meeting) and tags: as appropriate after import.pip install python-docx
Or use requirements.txt:
pip install -r requirements.txt
| Flag | Description |
|---|---|
-o, --output | Output markdown file path (default: same as docx with .md) |
-q, --quiet | Suppress progress messages |