Homelab MonitorBLOCK
Plug-and-play homelab dashboard in one container — GPU, local-AI VRAM, Docker, systemd, host health. Built-in read-only MCP server so AI agents can explore it too.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://github.com/SikamikanikoBG/homelab-monitor/stargazers) [](https://hub.docker.com/r/sikamikaniko123/homelab-monitor) [](https://discord.gg/tpKWKEdSQN) [](CHANGELOG.md) [](https://sikamikanikobg.github.io/homelab-monitor/)
One page for your whole home lab & AI rig — GPU truth (any vendor), tokens/sec, power cost by the hour, uptime, training runs, containers, disks. No agents, no separate metrics stack, no cloud.
Your home lab grew into a couple of machines, a Pi, and a GPU that's mysteriously always busy — and lately it's running models too. HomeLab Monitor gives you one self-hosted page that answers the real questions: what's that GPU actually doing, which model is holding it, what's it costing you to run, which container is eating RAM, what's filling your disks, and is anything down — across every box over SSH: Linux, a Pi, even Windows. Readable from your phone over the VPN.
Get started
# Grab the compose file and go. No GPU required — the GPU panels just light up when one's present. curl -fsSLO https://raw.githubuserc
395415fd9265OBSERVED · 2026-10-06Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add homelab-monitor:0.17.0 -- docker run -i --rm docker.io/sikamikaniko123/homelab-monitor:0.17.0:None python /app/mcp_server.py
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (14 observation(s))
- Network
- declared (8 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (25)
SSH_KEY = os.path.join(SSH_DIR, "id_ed25519")
"cmd": f"mkdir -p ~/.ssh && chmod 700 ~/.ssh\n"
f"echo '{key}' >> ~/.ssh/authorized_keys\n"f"chmod 600 ~/.ssh/authorized_keys",
"cmd": f"mkdir -p ~/.ssh && chmod 700 ~/.ssh\n"
CMD curl -fsS "http://127.0.0.1:${PORT:-9800}/healthz" || exit 1health = "http://127.0.0.1:%d/api/health" % port
COPILOT_OLLAMA_URL = os.environ.get("COPILOT_OLLAMA_URL", "http://127.0.0.1:11434").rstrip("/")# identically (host networking, GPU PID->container mapping, ptrace ports).
# Docker's default AppArmor profile (`docker-default`) denies ptrace
out = (out or "").lstrip("").strip()# private cgroup namespace where those paths are virtualised (0::/../../...),
self.assertNotIn("..", (app._safe_scan_path("/a/b/../../c") or "").split("/"))j = self.c.get("/api/disk_scan?path=../../etc&host=vader").get_json()if curl -fsS http://127.0.0.1:9800/healthz >/dev/null; then
if curl -fsS http://127.0.0.1:9800/healthz >/dev/null; then
mcp
mcp
- **A custom server on a remote host showed "0 loaded" while visibly serving traffic.** Every probe except Ollama's reports `vram=None` for every model it lists — there's no on-disk/loaded split to re
- **Benchmark Lab — a new AI tab that measures what your local models actually do on your GPU, and remembers it.** Pick one or more ollama models and a set of context sizes; the Lab loads each model a
Detect by listening port (8888/8889/8890) + `/proc/<pid>/cmdline` containing `jupyter-lab`/`jupyter-notebook`/`jupyter-server`. The REST API (`GET /api/sessions`, `/api/kernels`, `/api/status`) lists
design/brand/homelab-monitor-banner-1500.png
docs/demo.gif
docs/demo.mp4
Gates applied: no_behavioural_pass.
395415fd9265full audit observations/trust-audit/mcp-server/sikamikanikobg__homelab-monitor.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-06 | 395415fd9265 | BLOCK | D | 69 | first audit |
Questions
What is the Homelab Monitor MCP server?
Plug-and-play homelab dashboard in one container — GPU, local-AI VRAM, Docker, systemd, host health. Built-in read-only MCP server so AI agents can explore it too.
Is Homelab Monitor safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (69/100) and found 5 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What credentials does Homelab Monitor need?
It reads GH_TOKEN, GITHUB_TOKEN and HOMELAB_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Homelab Monitor run?
It speaks sse, stdio and streamable-http, so it runs as a local process your client starts.
How current is this page?
The grade is for one exact copy of the source (395415fd9265), read on 2026-10-06. The repository is watched and re-audited when it changes.