A local indexing tool that uses BM25 and vector embeddings to search markdown collections efficiently while minimizing token usage.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install anshumanbh-qmd
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 anshumanbh-qmd 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 Search?
QMD Search is a powerful utility designed for developers and researchers who maintain extensive markdown-based knowledge bases. By leveraging the qmd indexing engine, this skill allows AI agents to perform high-speed, local searches across directories like Obsidian vaults or documentation repositories. Unlike standard search methods that might ingest entire files, this integration for Openclaw Skills focuses on retrieving only the most relevant snippets, ensuring privacy and significant cost savings.
The tool utilizes a hybrid approach, combining traditional BM25 keyword matching with modern vector embeddings. This allows for both precise technical lookups and broad conceptual queries, making it an essential component of a local-first AI development workflow.
QMD Search Use Cases
- Searching through massive Obsidian vaults to find specific project context without hitting token limits.
- Finding relevant documentation snippets within a local markdown collection for coding assistance.
- Performing semantic searches for concepts where the exact terminology might be forgotten.
- Reducing LLM costs by passing highly relevant text fragments instead of reading thousands of full files.
How QMD Search Works
- The skill parses user input to identify the search query and determines whether to use keyword, semantic, or hybrid search.
- It verifies the local installation of the qmd CLI tool and checks for available collections.
- A search command is executed against the local index (using BM25 for keywords or vector similarity for concepts).
- The skill returns a list of highly relevant snippets accompanied by their respective file paths.
- The agent can then use these snippets to provide answers or decide to read specific files in their entirety if more context is needed.
QMD Search Setup
To get started with this skill, ensure you have the qmd utility installed and your collections indexed:
# Install the qmd utility
bun install -g https://github.com/tobi/qmd
# Add your markdown folder or Obsidian vault as a collection
qmd collection add ~/path/to/your/notes --name research
# Generate vectors for semantic search capability
qmd embed --collection research
QMD Search Data Schema & Taxonomy
The skill manages data through locally stored indexes and metadata. Use the following structure to understand how information is organized:
| Data Component | Description |
|---|---|
| Collections | Named groups of markdown files (e.g., 'work', 'notes'). |
| BM25 Index | A local keyword-based index used for fast, literal term matching. |
| Vector Store | Local embeddings that enable semantic similarity search for concepts. |
| Snippets | Extracted text fragments provided as output to maximize token efficiency. |
QMD Search Advanced Features
- Hybrid Search Mode: Combines BM25 and Vector search with LLM reranking for maximum accuracy.
- Semantic Vector Search: Enables conceptual discovery using the --semantic flag within Openclaw Skills.
- Multi-Collection Support: Seamlessly switch between different markdown repositories using the --collection argument.
- Local-First Privacy: All indexing and search operations occur entirely on your local machine.
SKILL.md
Loading
Related Openclaw Skills
Wheels Router Transit Planner
A high-performance transit routing skill for AI agents providing global trip planning with deep integration for Hong Kong transportation.
Jina DeepSearch
A high-performance deep search skill for AI agents to retrieve grounded information using Jina AI's advanced search models.
Training Manager
A comprehensive workspace orchestration tool for scaffolding, maintaining, and optimizing the training lifecycle of AI agents.
TrustLog Guard
A privacy-first financial governance tool that monitors OpenClaw API spending, enforces budgets, and detects cost anomalies locally.
SecureVibes Scanner
An AI-native security orchestration platform that uses Claude AI to perform deep application security scans, threat modeling, and continuous monitoring.
Hostex API Integration
A comprehensive Node.js integration for the Hostex OpenAPI to manage vacation rental properties, reservations, and guest communication.