Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
/plugin marketplace add dinglebear-ai/unraidnpx agentmods add plugins/dinglebear-ai/unraid/unraid-rsgit clone --depth 1 https://github.com/dinglebear-ai/unraidGrade A, and why
runraid scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
{
"name": "runraid",
"version": "0.5.0",
"description": "Unraid NAS/homelab monitoring via MCP",
"author": {
"name": "jmagar"
},
"repository": "https://github.com/dinglebear-ai/unraid",
"license": "MIT",
"keywords": [
"unraid",
"nas",
"homelab",
"mcp"
],
"skills": "./skills",
"userConfig": {
"server_url": {
"type": "string",
"title": "Server URL",
"description": "Base URL the MCP client connects to. Keep the default http://localhost:40010 when Claude Code runs on the same machine as the unraid-rmcp server. Change to e.g. http://nas:40010 or https://unraid.example.com when connecting to a remote instance. Do NOT append /mcp — the plugin adds it.",
"required": true,
"default": "http://localhost:40010"
},
"api_token": {
"type": "string",
"title": "API token",
"description": "Bearer token for MCP HTTP authentication (Authorization: Bearer <token>). Must match UNRAID_RMCP_TOKEN on the server. Generate one with: openssl rand -hex 32",
"sensitive": true,
"required": false
},
"no_auth": {
"type": "boolean",
"title": "Disable service auth",
"description": "Run unraid-rmcp without MCP auth. Use only when an upstream gateway enforces auth before traffic reaches this service.",
"required": true,
"default": false
},
"enabled_tools": {
"type": "string",
"title": "Enabled MCP tools/actions",
"description": "Optional comma-separated allowlist. Empty inherits the server's persisted .env/config policy; when none exists, all actions are enabled. Use action names such as array,docker,status,help or qualified selectors such as unraid.array.",
"default": ""
},
"disabled_tools": {
"type": "string",
"title": "Disabled MCP tools/actions",
"description": "Optional comma-separated denylist. Empty inherits the server's persisted .env/config policy. Deny rules override enabled rules. Use What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 120 lines scan A d952805ecc4e
runraid is a plugin published in the GitHub repository dinglebear-ai/unraid (129 stars, last pushed 2d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other plugins, from other repositories
cortex
Self-hosted log aggregation and investigation for homelabs: syslog, Docker, OTLP, and AI transcripts in SQLite/FTS, exposed over MCP, CLI, and REST.
sound
Code taste as installable rules. Probe a repo's stack, select only the rules its code shows a surface for (evidence-cited per rule), confirm, and scaffold a browsable .claude/sound/ rule map (surfaced by the sound:prime skill). Installed rules are user-owned; re-runs reconcile.
smk
Plugin marketplace listing 13 plugins: cursor-delegate, spring-clean, portal-skills, readable, decode.
zulipchat
Connect Claude Code sessions to Zulip topics with lifecycle hooks, owner approvals, and bidirectional control.
portal-skills
Platform engineering and security skills: GitLab CI/CD + Helm chart generation for Kubernetes, data-driven pod rightsizing from Prometheus, a proactive secure-coding baseline, and a deep OWASP reference (Top 10:2025, LLM Top 10, Agentic AI).
debugger
Debugging specialist for errors, test failures, and unexpected behavior. Use proactively when encountering any issues.