← Back to Skills Marketplace
Qmd
by
lifecoacher
· GitHub ↗
· v0.1.0
2187
Downloads
1
Stars
4
Active Installs
1
Versions
Install in OpenClaw
/install qmd-skill-2
Description
Local hybrid search for markdown notes and docs. Use when searching notes, finding related content, or retrieving documents from indexed collections.
Usage Guidance
This skill appears coherent for running a local Markdown search tool, but review a few points before installing: 1) The agent will call your local qmd binary — ensure you trust the qmd binary's source (the SKILL suggests installing via bun from a GitHub repo). 2) qmd can auto-download large GGUF models to ~/.cache/qmd/models (or XDG_CACHE_HOME); consider redirecting the cache or blocking network access if you don't want automatic model downloads. 3) Indexing is explicit (you must run qmd collection add /path/to/notes), but be careful not to index sensitive directories (password stores, system configs, private keys). 4) Installing via bun will fetch and install code from GitHub — run that install in a controlled environment or inspect the upstream repo first. If you want extra safety, run qmd in a sandboxed environment or review the upstream project (https://github.com/tobi/qmd) before granting the agent the ability to invoke it.
Capability Analysis
Type: OpenClaw Skill
Name: qmd-skill-2
Version: 0.1.0
The skill bundle is classified as benign. It transparently instructs the OpenClaw agent to install the 'qmd' local markdown search tool from its official GitHub repository via `bun install -g https://github.com/tobi/qmd` in SKILL.md. All subsequent instructions guide the agent on how to configure and use the 'qmd' tool for its stated purpose of local markdown search, including examples for maintenance via cron. There is no evidence of data exfiltration, unauthorized remote execution, persistence mechanisms beyond local maintenance, or prompt injection attempts to subvert the agent's security or purpose. The network and shell access are directly tied to the legitimate installation and operation of the specified tool.
Capability Assessment
Purpose & Capability
The name/description (local hybrid search for markdown notes) matches the runtime instructions: all commands reference the qmd CLI, indexing local files, and using local models. Required binaries (qmd) and suggested Bun install are appropriate for this purpose.
Instruction Scope
SKILL.md instructs the agent to run qmd commands (search, vsearch, query, embed, update, collection add/get) and to index/search files on disk — exactly what a local search tool must do. It does not ask for unrelated environment variables, system credentials, or to exfiltrate data to unexpected endpoints. It does note that qmd may auto-download GGUF models and stores them in ~/.cache/qmd/models (or XDG_CACHE_HOME).
Install Mechanism
The registry lists no formal install spec, but SKILL.md metadata includes an install suggestion: `bun install -g https://github.com/tobi/qmd`. Installing with bun from a GitHub repo is a common approach, but it involves fetching and executing remote code. This is proportionate to installing the qmd CLI but worth noting because it causes network fetches and code to be written/executed locally.
Credentials
No credentials or sensitive env vars are requested. The only environment guidance is PATH ($HOME/.bun/bin) and an optional XDG_CACHE_HOME override for model cache. Those are appropriate for tool installation and cache placement.
Persistence & Privilege
The skill does not request always:true and will not be force-enabled across agents. The instructions include cron/scheduler examples to keep indexes fresh (user-controlled) but do not modify other skills or system-wide agent configs. Note: autonomous invocation is allowed by platform default, so an agent could run qmd commands on its own when invoked.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install qmd-skill-2 - After installation, invoke the skill by name or use
/qmd-skill-2 - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
Initial release of qmd-skill-2: local hybrid search for Markdown notes and docs.
- Provides fast keyword search (`qmd search`) and slow semantic search (`qmd vsearch`) over indexed Markdown collections.
- Highlights best usage practices, install and setup instructions, and behavior differences between keyword and semantic search modes.
- Lists common commands and important options for searching, retrieving, and maintaining your Markdown search index.
- Offers advice on keeping indexes updated and embedding refreshed via cron or scheduler.
- Explains relation to other search tools (e.g., code search, Clawdbot memory search).
- Requires Bun and is compatible with macOS and Linux.
Metadata
Frequently Asked Questions
What is Qmd?
Local hybrid search for markdown notes and docs. Use when searching notes, finding related content, or retrieving documents from indexed collections. It is an AI Agent Skill for Claude Code / OpenClaw, with 2187 downloads so far.
How do I install Qmd?
Run "/install qmd-skill-2" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Qmd free?
Yes, Qmd is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Qmd support?
Qmd is cross-platform and runs anywhere OpenClaw / Claude Code is available (darwin, linux).
Who created Qmd?
It is built and maintained by lifecoacher (@lifecoacher); the current version is v0.1.0.
More Skills