anshumanbh-qmd
от anshumanbh
Search & ResearchКратко
Search markdown knowledge bases efficiently.
Описание
Что делает этот навык
qmd indexes local markdown files and returns relevant snippets using BM25 keyword matching and vector embeddings. It targets Obsidian vaults and similar markdown collections. Searches return file paths and context excerpts rather than full file contents.
Returns only relevant snippets instead of full files, cutting token usage by roughly 96% compared to reading markdown files directly into the context window.
Когда его использовать
- Finding a specific note in a large Obsidian vault
- Searching personal knowledge bases by concept
- Locating technical documentation by keyword
- Retrieving past meeting notes or project decisions
- Querying a markdown journal without reading every file
Пример сценария
User asks: find my notes on API authentication
- 1
Parse 'api authentication' as search query from arguments
- 2
Check that qmd is installed with 'which qmd'
- 3
List available collections with 'qmd collection list'
- 4
Run 'qmd search "api authentication" --collection notes'
- 5
Present returned snippets with file paths to the user
Relevant snippet excerpts from matching notes, each with a file path for direct reference
Требования
- qmd installed via Bun from github.com/tobi/qmd
- Bun runtime (required to install qmd globally)
Установка
clawhub install anshumanbh/anshumanbh-qmdВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.