archive/skills/qmd-knowledge/SKILL.md
Knowledge management with QMD (Quick Markdown Search). Search, retrieve, and organize your personal knowledge base, notes, docs, and meeting transcripts. Use when asking about past notes, documentation, meetings, or when needing to recall information from your indexed markdown files.
npx skillsauth add michalvavra/agents qmd-knowledgeInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
QMD (Quick Markdown Search) is an on-device search engine for your knowledge base. It indexes markdown notes, meeting transcripts, documentation, and more, providing fast keyword search, semantic search, and hybrid search with LLM reranking.
| Tool | Best For | Speed |
|------|----------|-------|
| qmd_search | Exact keywords, known terms | Fast |
| qmd_vsearch | Conceptual queries, fuzzy matching | Medium |
| qmd_query | Best quality, complex questions | Slower |
| qmd_get | Retrieve full document content | Fast |
| qmd_multi_get | Get multiple docs at once | Fast |
| qmd_status | Check index health | Fast |
Use qmd_search when:
Use qmd_vsearch when:
Use qmd_query when:
After searching, use qmd_get with the path or docid to fetch full content:
# By path (from search results)
qmd_get { "path": "notes/meeting-2024-01-15.md" }
# By docid (the #xxx from results)
qmd_get { "path": "#abc123" }
# Multiple documents
qmd_multi_get { "pattern": "docs/api/*.md" }
collection: "notes")minScoreqmd_queryqmd_search { "query": "quarterly planning 2024", "collection": "meetings" }
qmd_vsearch { "query": "how to set up CI/CD pipeline" }
qmd_query { "query": "authentication flow for API users" }
qmd_get { "path": "#a1b2c3", "full": true }
Search results include:
qmd_get #xxx)Users organize content into collections. Check what's available:
qmd_status {}
Common collections:
notes - Personal notesdocs - Documentationmeetings - Meeting transcriptswork - Work-related contentFilter searches by collection when user's intent is clear.
Start broad, then narrow: Begin with qmd_query if unsure, refine with filters
Combine tools: Search first, then qmd_get to retrieve full content
Use appropriate search type:
qmd_searchqmd_vsearchqmd_queryCheck status first: If searches fail, qmd_status can reveal issues
Respect user's collections: Don't search everywhere if user mentions specific notes
"No results found":
qmd_vsearch)minScoreqmd_status for collection names"Embeddings not ready":
qmd embedqmd_search still works (BM25 only)"Command not found":
brew install oven-sh/bun/bun
bun install -g https://github.com/tobi/qmd
# Install Bun (required runtime for QMD)
brew install oven-sh/bun/bun
# Install QMD
bun install -g https://github.com/tobi/qmd
# Add collections
qmd collection add ~/notes --name notes
qmd collection add ~/Documents/meetings --name meetings
# Generate embeddings (for semantic search)
qmd embed
# Keyword search
qmd search "API authentication"
# Semantic search
qmd vsearch "how to log in"
# Hybrid (best quality)
qmd query "user auth flow"
# Get document
qmd get "docs/api.md"
qmd get "#abc123"
# Multiple documents
qmd multi-get "journals/2025-05*.md"
# Status
qmd status
devops
Use when preparing, reviewing, or documenting versioned app releases, changelogs, SemVer bumps, release commits, git tags, or tag-gated production deploys
development
Staff-level codebase health review. Finds monolithic modules, silent failures, type safety gaps, test coverage holes, and LLM-friendliness issues.
documentation
Write AI-scannable technical documentation.
tools
CLI for Snowflake. Query data, manage warehouses, databases, schemas, tables, and stages. Use when working with Snowflake data platform.