Files
unraid-mcp/skills/unraid/README.md
Jacob Magar 4a694c0f32 docs: fix markdown lint, broken links, stale counts, and publishing guidance
- Fix broken ToC anchors in competitive-analysis.md (MD051)
- Add blank lines before code blocks in api-reference.md (MD031)
- Add language identifiers to directory tree code blocks in MARKETPLACE.md and skills/unraid/README.md (MD040)
- Fix size unit guidance conflict: clarify disk sizes are KB, memory is bytes
- Update stale "90 actions" references to "76 actions" across research docs
- Fix coverage table terminology and clarify 22% coverage calculation
- Recommend PyPI Trusted Publishing (OIDC) over API token secrets in PUBLISHING.md
- Update action count in .claude-plugin/README.md

Resolves review threads: PRRT_kwDOO6Hdxs5uvO2m, PRRT_kwDOO6Hdxs5uvO2o,
PRRT_kwDOO6Hdxs5uvO2r, PRRT_kwDOO6Hdxs5uvOcl, PRRT_kwDOO6Hdxs5uvOcr,
PRRT_kwDOO6Hdxs5uvKrq, PRRT_kwDOO6Hdxs5uvO2u, PRRT_kwDOO6Hdxs5uvO2w,
PRRT_kwDOO6Hdxs5uvO2z, PRRT_kwDOO6Hdxs5uu7zl
2026-02-15 23:05:05 -05:00

161 lines
3.8 KiB
Markdown

# Unraid API Skill
Query and monitor Unraid servers via the GraphQL API.
## What's Included
This skill provides complete access to all 27 read-only Unraid GraphQL API endpoints.
### Files
```text
skills/unraid/
├── SKILL.md # Main skill documentation
├── README.md # This file
├── scripts/
│ └── unraid-query.sh # GraphQL query helper script
├── examples/
│ ├── monitoring-dashboard.sh # Complete system dashboard
│ ├── disk-health.sh # Disk temperature & health check
│ └── read-logs.sh # Log file reader
└── references/
├── api-reference.md # Complete API documentation
└── quick-reference.md # Common queries cheat sheet
```
## Installation
This skill is part of the Unraid MCP plugin. Install via the Claude Code marketplace:
```bash
/plugin marketplace add jmagar/unraid-mcp
/plugin install unraid @unraid-mcp
```
The plugin includes both the MCP server and this skill at `skills/unraid/`.
## Quick Start
1. **Set your credentials:**
```bash
export UNRAID_URL="https://your-unraid-server/graphql"
export UNRAID_API_KEY="your-api-key"
```
2. **Run a query:**
```bash
cd skills/unraid
./scripts/unraid-query.sh -q "{ online }"
```
3. **Run examples:**
```bash
./examples/monitoring-dashboard.sh
./examples/disk-health.sh
```
## Triggers
This skill activates when you mention:
- "check Unraid"
- "monitor Unraid"
- "Unraid API"
- "Unraid disk temperatures"
- "Unraid array status"
- "read Unraid logs"
- And more Unraid-related monitoring tasks
## Features
- **27 working endpoints** - All read-only queries documented
- **Helper script** - Easy CLI interface for GraphQL queries
- **Example scripts** - Ready-to-use monitoring scripts
- **Complete reference** - Detailed documentation with examples
- **Quick reference** - Common queries cheat sheet
## Endpoints Covered
### System & Monitoring
- System info (CPU, OS, hardware)
- Real-time metrics (CPU %, memory %)
- Configuration & settings
- Log files (list & read)
### Storage
- Array status & disks
- All physical disks (including cache/USB)
- Network shares
- Parity check status
### Virtualization
- Docker containers
- Virtual machines
### Power & Alerts
- UPS devices
- System notifications
### Administration
- API key management
- User & authentication
- Server registration
- UI customization
## Requirements
- **Unraid 7.2+** (GraphQL API)
- **API Key** with Viewer role
- **jq** for JSON parsing (usually pre-installed)
- **curl** for HTTP requests
## Getting an API Key
1. Log in to Unraid WebGUI
2. Settings → Management Access → API Keys
3. Click "Create API Key"
4. Name: "monitoring" (or whatever you like)
5. Role: Select "Viewer" (read-only)
6. Copy the generated key
## Documentation
- **SKILL.md** - Start here for task-oriented guidance
- **references/api-reference.md** - Complete endpoint reference
- **references/quick-reference.md** - Quick query examples
## Examples
### System Status
```bash
./scripts/unraid-query.sh -q "{ online metrics { cpu { percentTotal } } }"
```
### Disk Health
```bash
./examples/disk-health.sh
```
### Complete Dashboard
```bash
./examples/monitoring-dashboard.sh
```
### Read Logs
```bash
./examples/read-logs.sh syslog 20
```
## Notes
- Disk/array sizes are in **kilobytes**; memory values (from `info.memory` and `metrics.memory`) are in **bytes**
- Temperatures are in **Celsius**
- Docker container logs are **not accessible** via API (use SSH)
- Poll no faster than every **5 seconds** to avoid server load
## Version
- **Skill Version:** 0.2.0
- **API Version:** Unraid 7.2 GraphQL
- **Tested:** 2026-01-21
- **Endpoints:** 27 working read-only queries