qmd

Name: qmd
Availability: InStock
Author: joeleung0122

Index local files and search them with BM25-based ranking.

Updated Feb 3, 2026

Search Tags:#indexing #cli #search #local #vectors #rerank #bm25

Ready to run

npx skills add https://github.com/joeleung0122/testing --skill qmd-joeleung0122

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: qmd
Source: https://github.com/joeleung0122/testing/tree/main/skills/qmd
Command: npx skills add https://github.com/joeleung0122/testing --skill qmd-joeleung0122

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Indexing local files to enable fast, accurate search across personal or project data without relocating data to a central service.

Core Features & Use Cases

Local indexing: Build BM25 indexes and vector-backed results for quick retrieval.
Hybrid search: Combine BM25 ranking with vector similarity and optional rerank to improve relevance.
Use Case: Quickly locate a code snippet or document across a local workspace using natural language queries.

Quick Start

Run the qmd CLI to index your documents and perform a search.

Frequently Asked Questions about qmd

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

FAQPage Schema

How do I search local files using BM25 and vector similarity?▼

You can search local files using BM25 and vector similarity by indexing your personal data stores or developer projects with a CLI tool. This builds local indexes for quick retrieval without relocating data to a central service.

What is the best way to find a specific code snippet across a local workspace?▼

The best way to find a code snippet across a local workspace is using a hybrid search approach. It combines BM25 ranking with vector similarity and optional reranking to improve relevance for natural language queries.

Can I build a local search index for my personal knowledge base without uploading data?▼

Yes, you can build a local search index for a personal knowledge base without uploading data. The tool runs as a CLI on your device, indexing local files to enable on-device search tasks while keeping data private.

Does hybrid search with BM25 and reranking improve document retrieval relevance?▼

Hybrid search with BM25 and reranking improves document retrieval relevance by combining traditional keyword-based ranking with vector similarity. Applying optional reranking further refines the results for more accurate local file searches.

When should I use a local CLI search tool instead of a centralized search service?▼

You should use a local CLI search tool instead of a centralized search service when you need on-device search tasks for personal data stores or developer projects. It allows you to index local files for fast retrieval without relocating data.

Why does qmd combine BM25 indexing with vector search for local files?▼

qmd combines BM25 indexing with vector search to deliver accurate local search results. BM25 handles keyword matching while vector search captures semantic similarity, and optional reranking optimizes the final order.