← Back to Skills Marketplace
emaspa

UnraidClaw

by emaspa · GitHub ↗ · v0.1.9 · MIT-0
cross-platform ⚠ suspicious
396
Downloads
0
Stars
2
Active Installs
6
Versions
Install in OpenClaw
/install unraidclaw
Description
Manage your Unraid server through AI agents - 43 tools for Docker, VMs, array, shares, system, notifications, and more with permission control.
Usage Guidance
This plugin appears to do what it says: it requires you to supply a UnraidClaw server URL and API key in your OpenClaw plugin config. Before installing: 1) Verify the package source (the registry metadata showed no homepage/source — confirm the GitHub/npm links in SKILL.md are correct and match the package you install). 2) Treat the API key as sensitive — store it only in the plugin config and rotate/revoke if compromised. 3) Avoid enabling tlsSkipVerify unless you must for a local self-signed cert (it disables TLS verification and increases MITM risk). 4) Limit which agents or users can call the plugin, or require human approval for destructive actions (reboot, shutdown, array stop, remove containers, create containers with host mounts), since the plugin exposes those capabilities. 5) If you have strict security requirements, review the source (the included files are readable) and test in a non-production environment first.
Capability Analysis
Type: OpenClaw Skill Name: unraidclaw Version: 0.1.9 The unraidclaw skill provides a comprehensive suite of 43 tools for remote Unraid server management, including Docker, VM, and system controls. It uses a standard HTTP/HTTPS client (src/client.ts) to communicate with a user-configured server using an API key. While it possesses high-privilege capabilities (e.g., unraid_docker_create, unraid_system_reboot) and allows bypassing TLS verification (tlsSkipVerify), these features are documented and aligned with its administrative purpose. No evidence of malicious intent, obfuscation, or unauthorized data exfiltration was found.
Capability Assessment
Purpose & Capability
Name/description describe Unraid management and the code registers tools for Docker, VMs, array, shares, system, notifications, network, users, logs, and health. The code constructs HTTP requests to a configured UnraidClaw server and requires an API key and server URL via plugin config — exactly what you'd expect for this functionality.
Instruction Scope
SKILL.md and source instruct installation as an OpenClaw plugin and storing serverUrl/apiKey in openclaw.json. Runtime code reads only plugin/openclaw config and uses it to call the UnraidClaw API. There are destructive operations (reboot, shutdown, stop array, remove containers, create containers with volumes/envs) but these are in-scope for Unraid management and are expected; the skill does not instruct reading unrelated system files or environment variables.
Install Mechanism
The registry lists no automated install spec; SKILL.md instructs installing the npm package using npm pack and openclaw plugins install, which is a normal local plugin install flow. The package.json references the dist entrypoint and an npm package. Minor inconsistency: package.json shows version 0.1.6 while registry metadata states 0.1.9; source/homepage were reported as 'unknown' / 'none' in registry metadata despite SKILL.md including GitHub/npm links — verify you are installing the intended package from the expected source.
Credentials
No environment variables are required by the registry; the plugin expects an API key and serverUrl in the OpenClaw plugin config (openclaw.plugin.json schema and SKILL.md). This is proportionate to remote Unraid API access. The only additional option is tlsSkipVerify for self-signed certs — legitimate but reduces TLS security if enabled.
Persistence & Privilege
always:false (default) and model invocation is allowed (default). The plugin registers many management tools including destructive actions; autonomous agent invocation could therefore perform destructive operations if policies permit. This is expected for a management plugin, but operators should control which agents/flows are allowed to invoke high-impact tools.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install unraidclaw
  3. After installation, invoke the skill by name or use /unraidclaw
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.9
Fix install instructions: use npm pack workaround instead of non-existent --registry flag
v0.1.8
Fix install instructions: use --registry npm to install as plugin, not skill
v0.1.6
Add multi-server support: configure multiple Unraid servers, target with optional server parameter
v0.1.5
Add memory, CPU load, and disk usage stats to tool descriptions
v0.1.4
Add SKILL.md frontmatter for ClawHub metadata
v0.1.3
- Added comprehensive SKILL.md documentation detailing all UnraidClaw features and configuration. - Outlined available tools (43 tools in 11 categories) and permission controls. - Listed requirements, installation steps, configuration options, and usage examples. - Provided relevant resource links (GitHub, npm, Unraid Community Apps).
Metadata
Slug unraidclaw
Version 0.1.9
License MIT-0
All-time Installs 2
Active Installs 2
Total Versions 6
Frequently Asked Questions

What is UnraidClaw?

Manage your Unraid server through AI agents - 43 tools for Docker, VMs, array, shares, system, notifications, and more with permission control. It is an AI Agent Skill for Claude Code / OpenClaw, with 396 downloads so far.

How do I install UnraidClaw?

Run "/install unraidclaw" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is UnraidClaw free?

Yes, UnraidClaw is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does UnraidClaw support?

UnraidClaw is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created UnraidClaw?

It is built and maintained by emaspa (@emaspa); the current version is v0.1.9.

💬 Comments