PLAN: the KVM is a Sipeed NanoKVM at 10.0.0.217 — framing is fine, HTTP is the blocker

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

PLAN.md

@@ -553,31 +553,37 @@ something long in one, switch away and back, refresh — it's all still there.
   item names the heavier pieces (multi-select and bulk actions, folder download, resumable
   uploads, a gallery, search) — worth doing the two together, since they are the same tab.
 
-- **A KVM tab** — asked for 2026-09-24, not started. A tab that shows a **custom IP-KVM's
+- **A KVM tab** — asked for 2026-09-24, not started. A tab that shows the **IP-KVM's own web
-  own web page** inside Ham Lab, so the machine's screen, keyboard and mouse are one click
+  page** inside Ham Lab, so a machine's screen, keyboard and mouse are one click away. **This
-  away. **This is KVM the keyboard-video-mouse switch, not KVM the hypervisor** — the
+  is KVM the keyboard-video-mouse switch, not KVM the hypervisor** — the Virtualization item
-  Virtualization item above means the other one, and the two must not be confused when
+  above means the other one, and the two must not be confused when either is built.
-  either is built. No such device exists in the lab yet, so this starts with the hardware
+  - **The device**: a **Sipeed NanoKVM**, attached to `grid`, at **`10.0.0.217`** (MAC
-  (PiKVM, JetKVM, TinyPilot, or the home-built one Mitchell has in mind).
+    `48:da:35:6f:d0:31`), serving a web UI on port 80. The only one in the lab so far, so the
-  - **Where the entry lives is the design decision.** A KVM earns its keep exactly when the
+    tab must be built for a list of them, not hard-wired to one.
-    machine is *down* — no OS, a bad kernel, a BIOS screen — which is when that host's Ham
+  - **Framing works.** Probed 2026-09-24: the NanoKVM sends **no `X-Frame-Options` and no
-    Lab cannot answer. So the tab has to be **hub-side**, a list the hub holds, and it must
+    `frame-ancestors`**, so unlike most of its competitors it does not refuse to be embedded.
-    work for machines that run no Ham Lab at all. Configuring it as a `kvm_url` inside each
+    An iframe is a real option and the fallback of "just link out" is not needed.
-    node's own `config.json` would fail in precisely the case it is for.
+  - **Mixed content is the actual blocker.** The NanoKVM is **HTTP only** — there is no
-  - **Framing is the first thing to test, before any UI is written.** Most IP-KVMs send
+    HTTPS listener at all — and Ham Lab is served over TLS, so a browser will refuse the
-    `X-Frame-Options: DENY` or `frame-ancestors 'none'` and simply refuse to be embedded —
+    iframe outright. Something must give it HTTPS. Two routes:
-    Ham Lab sends both itself, on its own pages. A device whose nginx can be edited (PiKVM)
+    - **Proxy it through the daemon**, which is the better one. Ham Lab already
-    or that is home-built can be told to allow it; one that cannot will show a blank box,
+      reverse-proxies nodes over WebSockets (`proxyNode`), which is exactly what the KVM's
-    and then the honest answer is a link that opens it in a tab rather than a fake embed.
+      video and input streams need, and it puts the KVM behind **Ham Lab's own login** rather
-  - **Mixed content and certificates.** Ham Lab is served over HTTPS through NPM. A KVM on
+      than a second one. `/api/vm/info` answers **401**, so the device has its own auth; the
-    plain HTTP will be blocked by the browser outright, and one with a self-signed
+      hub could hold that credential once instead of prompting per person.
-    certificate fails silently inside an iframe. Either give it a proper cert and an NPM
+    - **Or an NPM proxy host** (`scripts/npm-proxy-host.sh`), which is two minutes of work —
-    host, or proxy it through the daemon — which already reverse-proxies nodes and carries
+      but note what it means: `*.bigham.dev` is a wildcard to the public IP, so that would put
-    WebSockets (`proxyNode`), and a KVM's video and input streams are exactly that. Proxying
+      **BIOS-level keyboard and screen access to `grid` on the public internet** behind
-    also solves the separate login, since the credential could be held once on the hub.
+      whatever the KVM's own login is worth. If that route is taken it wants an NPM access
-  - Worth settling too: whether the tab is per machine (a machine's KVM shown on its page)
+      list or a LAN/WireGuard-only binding, decided deliberately.
-    or one page listing every KVM; and whether a machine with no Ham Lab at all deserves a
+  - **Where the entry lives is still the design decision.** A KVM earns its keep when the
-    first-class entry on the Machines page, which today only knows hosts that answer.
+    machine is *down* — no OS, a bad kernel, a BIOS screen — which is when that host's Ham Lab
+    cannot answer. So the list has to be **hub-side** and must work for machines running no
+    Ham Lab at all. A `kvm_url` inside each node's own `config.json` would fail in precisely
+    the case it exists for.
+  - Worth settling too: whether the KVM shows on its machine's page or on one page listing
+    every KVM; and whether a machine with no Ham Lab deserves a first-class entry on the
+    Machines page, which today only knows hosts that answer.
 
 ## Open questions