Bound a panel credential where the route check cannot reach

Three things the security pass on the portal pairing turned up. The first two
were already true of a screen on the local network; what changed is that a
panel credential is now presentable from the internet, which is what makes
them worth closing rather than recording.

The artifact endpoint authenticates for itself, because a worker's credential
has to open it and that token is no use anywhere else. It resolved the caller
without handing over the request, so the one credential that is scoped by
route was judged by no route at all — a panel could read and write the store
as whoever approved it. It passes the request it already holds now.

The websocket has no route to judge either, and there the bound has to be on
what is sent: a panel is given the values its own dashboards draw and nothing
else — no node status, no logs, no shape of the graph. The keys stay in the
message, emptied, because a screen on a wall runs the bundle it was paired
with. `messages_for` reads that set off the published dashboards, and is the
walk the `/messages/` allowlist has wanted for a while.

And locality is no longer a header anyone can type. The marker the connector
stamps is a value minted per process, so reaching this API directly cannot buy
a device the credential meant for one that cannot reach it at all.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017F9RnYCJgASuBTcAjxmnsp

NOTEPAD.md

@@ -186,7 +186,7 @@ as an em dash.
 - CHORE/UI: `ROW_HEIGHT` is a fixed 80px while column width follows the canvas, so a 1920-wide panel at 12 columns has 160×80 cells. If that reads too wide, the row height could derive from the canvas too.
 - CHORE/UI: multi-page and multi-section dashboards still have no UI, and now need none — a panel carries several whole dashboards instead, each with its own canvas and its own publish. `PageDef`/`SectionDef` stay in the schema and the editor still edits `sectionsOf(page)[0]`, so the page `Tabs` in `DashboardEditor` are dead until something writes a second page through the API.
 - FEAT/UI: a panel does not notice being reassigned until it is reloaded — nothing pushes the panel document or a dashboard publish, so the rail is as stale as the last read. Same gap as the wallpanel hot-reload item above; one event on the bus would answer both.
-- CHORE/API: a panel credential may publish *any* message, not only the ones its own widgets bind to — the allowlist is the `/messages/` prefix rather than a walk of the panel's widgets. Enough for a screen in a house; an installation where a panel sits somewhere less trusted would want the narrower check.
+- CHORE/API: a panel credential may publish *any* message, not only the ones its own widgets bind to — the allowlist is the `/messages/` prefix rather than a walk of the panel's widgets. The walk now exists: `panels.messages_for()` is what bounds the live socket. Pointing `_panel_may` at it would close this too, but it tightens what already-paired screens may do, so it wants a deliberate look at the query-chart request path first.
 - CHORE/API: unpairing a device means deleting the panel. A per-panel nonce in the token, bumped on demand, would let one screen be re-paired without disturbing the assignment.
 - CHORE/API: a panel paired through the portal is revoked here the moment the panel is deleted — `_panel_may` finds nothing and answers 401 — but the hub's copy of the token stays valid until it expires or the installation's generation counter is bumped ("New code"). The hub has no per-panel revocation, and giving it one means telling it which panels exist, which is exactly what this design avoids. The generation bump is the lever; it is blunt, cutting every credential the portal minted for the installation.
 - CHORE/UI: the device line under a pairing code is the raw user agent plus the address the request came from. Both are self-reported and neither is proof; it is there so an admin can tell the screen they just hung from one they were not expecting, not to authenticate anything.