qmd

Search markdown and text documents with hybrid keyword and semantic retrieval.

Updated May 5, 2026
One-click install
npx skills add https://github.com/Z43L/zeus-agent --skill qmd-z43l
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: qmd
Source: https://github.com/Z43L/zeus-agent/tree/main/optional-skills/research/qmd
Command: npx skills add https://github.com/Z43L/zeus-agent --skill qmd-z43l

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Finding the right information inside a growing collection of personal notes, meeting transcripts, and documents becomes slow and unreliable with simple keyword search.

Core Features & Use Cases

  • Hybrid retrieval (BM25 + vector search): Combines fast lexical matching with semantic retrieval for better recall.
  • LLM reranking for higher-quality results: Improves ordering of the most relevant matches using reranker models.
  • Local indexing and MCP/CLI access: Runs on-device without cloud dependencies, and can expose tools via MCP to integrate with Zeus Agent.
  • Use cases: “What did we decide about X?”, “Find notes about Y concept,” “Search across projects and meetings,” and “Retrieve exact documents once you have a match.”

Quick Start

Ask your agent to set up qmd by querying your notes with hybrid search for the phrase “what decisions were made about the database migration”.

Frequently Asked Questions about qmd

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How does hybrid retrieval combine BM25 and vector search for local knowledge bases?

You can search markdown and text documents across personal knowledge bases using hybrid keyword and semantic retrieval, with optional LLM reranking to improve the ordering of the most relevant matches.

Can I run local document search without cloud dependencies?

Yes, local document search runs entirely on-device without cloud dependencies, performing local indexing and embeddings so your personal notes and meeting transcripts remain private while still supporting fast hybrid querying.

How do I integrate local vector RAG into my agent workflow?

You integrate local vector RAG into agent workflows by exposing search and retrieval tools via MCP, enabling your agent to query indexed markdown and text documents and retrieve exact documents by ID or path.

What is the best way to search meeting transcripts for past decisions?

The best way to search meeting transcripts for past decisions is using hybrid BM25 and vector retrieval, which combines fast keyword matching with semantic understanding to quickly find what was decided about specific topics in your notes.

How do I retrieve a specific document by ID after a hybrid search?

After a hybrid search, you can retrieve exact documents by querying their specific document ID or file path, enabling follow-up retrieval to pull the full source text directly from your locally indexed knowledge base.