A high-performance local search and indexing CLI that combines BM25 keyword matching with vector-based semantic search and MCP support.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install qmd-quality-markdown
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install qmd-quality-markdown using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is qmd?
qmd is a specialized command-line utility built for developers who need to index and search local files with high precision. By integrating traditional BM25 search with modern vector embeddings and reranking, it provides a robust solution for navigating large local datasets, documentation, or codebases. This tool is a valuable addition to the ecosystem of Openclaw Skills, enabling users to maintain a local knowledge base that is both searchable and accessible to AI agents.
The tool is designed for privacy and speed, utilizing local resources for both indexing and inference. It serves as a bridge between your raw files and intelligent retrieval systems, ensuring that you can find exactly what you need without relying on external cloud services.
qmd Use Cases
- Indexing and searching through extensive local markdown documentation.
- Performing hybrid semantic searches across large code repositories to find context.
- Enabling AI agents to access local file content via the Model Context Protocol (MCP).
- Extracting specific line ranges from indexed files for precise technical referencing.
How qmd Works
- Add a collection of local files to the index using a directory path and optional file masks.
- Execute an update command to process files and generate searchable indices.
- Perform a search using BM25 for keyword matching or vector search for semantic relevance.
- Use hybrid query modes to combine search methods and rerank results via Ollama.
- Retrieve specific document fragments or metadata based on search rankings.
qmd Setup
To get started with this tool within the Openclaw Skills framework, install it via npm:
npm install -g https://github.com/tobi/qmd
Ensure you have Ollama running locally for embeddings and reranking support:
# Default URL: http://localhost:11434
export OLLAMA_URL="http://localhost:11434"
Initialize your first document collection:
qmd collection add /path/to/docs --name docs --mask "**/*.md"
qmd update
qmd Data Schema & Taxonomy
qmd organizes its indexing data locally to ensure performance and privacy. The system tracks the following components:
| Component | Details |
|---|---|
| Cache Storage | Default location is ~/.cache/qmd. |
| Collections | Named groups of files with specific inclusion masks (e.g., **/*.md). |
| Index Types | Supports BM25 (keyword), Vector (embeddings), and Hybrid indices. |
| Metadata | Stores file paths, status, and line counts for granular retrieval. |
qmd Advanced Features
- MCP Mode: Launch a Model Context Protocol server using
qmd mcpfor seamless AI agent integration. - Hybrid Querying: Combine vector and keyword search results for superior accuracy.
- Local Reranking: Use local Ollama instances to rerank search results based on query context.
- Granular Retrieval: Fetch specific line ranges from documents to minimize token usage in agent workflows.
- Flexible Masking: Use glob patterns to precisely control which files are indexed in a collection.
SKILL.md
Loading
Related Openclaw Skills
qmd
A local search and indexing CLI that combines BM25, vector search, and reranking with native MCP support for AI agents.
Humanizer
A sophisticated text editing skill that identifies and removes robotic patterns to inject personality into AI-generated content.
Content Machine
An autonomous 24/7 affiliate content production pipeline that orchestrates research, writing, quality control, and deployment for niche sites.
Larry: Autonomous TikTok Slideshow Agent
An AI-driven agent that automates the research, creation, and posting of high-engagement TikTok slideshows for affiliate marketing.
Multi Search Engine
A comprehensive search integration providing API-less access to 17 global and domestic engines with advanced filtering capabilities.
Finance Search Agent Strategy
A specialized AI research agent designed to execute deep financial market analysis, social media trend tracking, and verified report generation.
Featured*
